graphql-engine/server/tests-py/conftest.py

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

987 lines
38 KiB
Python
Raw Normal View History

import collections
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
import http.server
import inspect
import os
2018-09-18 09:21:57 +03:00
import pytest
import re
import socket
import sqlalchemy
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
import sys
import threading
import time
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
from typing import Optional
import urllib.parse
import uuid
import auth_webhook_server
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
from context import ActionsWebhookServer, EvtsWebhookServer, GQLWsClient, GraphQLWSClient, HGECtx, HGECtxGQLServer, HGECtxWebhook, PytestConf
import fixtures.hge
import fixtures.postgres
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
import fixtures.tls
import ports
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
import webhook
2018-10-30 12:21:58 +03:00
2018-09-18 09:21:57 +03:00
def pytest_addoption(parser):
parser.addoption(
"--hge-bin",
metavar="HGE_BIN",
required=False,
help="Hasura GraphQL Engine binary executable",
)
2018-09-18 09:21:57 +03:00
parser.addoption(
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
"--hge-urls",
metavar="HGE_URLS",
help="csv list of urls for graphql-engine",
required=False,
nargs='+'
2018-09-18 09:21:57 +03:00
)
parser.addoption(
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
"--pg-urls", metavar="PG_URLS",
help="csv list of urls for connecting to Postgres directly",
required=False,
nargs='+'
2018-09-18 09:21:57 +03:00
)
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
parser.addoption('--tls-ca-cert', help='The CA certificate used for helper services', required=False)
parser.addoption('--tls-ca-key', help='The CA key used for helper services', required=False)
parser.addoption(
"--hge-jwt-key-file", metavar="HGE_JWT_KEY_FILE", help="File containing the private key used to encode jwt tokens using RS512 algorithm", required=False
)
parser.addoption(
"--hge-jwt-conf", metavar="HGE_JWT_CONF", help="The JWT conf", required=False
)
2018-09-18 09:21:57 +03:00
parser.addoption(
"--test-ws-init-cookie",
metavar="read|noread",
required=False,
help="Run testcases for testing cookie sending over websockets"
)
parser.addoption(
"--test-hge-scale-url",
metavar="<url>",
required=False,
help="Run testcases for horizontal scaling"
)
parser.addoption(
"--test-logging",
action="store_true",
default=False,
required=False,
help="Run testcases for logging"
)
parser.addoption(
"--test-startup-db-calls",
action="store_true",
default=False,
required=False,
help="Run testcases for startup database calls"
)
parser.addoption(
"--accept",
action="store_true",
default=False,
required=False,
help="Accept any failing test cases from YAML files as correct, and write the new files out to disk."
)
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
parser.addoption(
"--skip-schema-teardown",
action="store_true",
default=False,
required=False,
help="""
Skip tearing down the schema/Hasura metadata after tests. This option may result in test failures if the schema
has to change between the list of tests to be run
"""
)
parser.addoption(
"--skip-schema-setup",
action="store_true",
default=False,
required=False,
help="""
Skip setting up schema/Hasura metadata before tests.
This option may result in test failures if the schema has to change between the list of tests to be run
"""
)
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
parser.addoption(
"--avoid-error-message-checks",
action="store_true",
default=False,
required=False,
help="""
This option when set will ignore disparity in error messages between expected and response outputs.
Used basically in version upgrade/downgrade tests where the error messages may change
"""
)
parser.addoption(
"--collect-upgrade-tests-to-file",
metavar="<path>",
required=False,
help="When used along with collect-only, it will write the list of upgrade tests into the file specified"
)
2020-05-05 22:57:17 +03:00
parser.addoption(
"--test-unauthorized-role",
action="store_true",
help="Run testcases for unauthorized role",
)
parser.addoption(
"--test-no-cookie-and-unauth-role",
action="store_true",
help="Run testcases for no unauthorized role and no cookie jwt header set (cookie auth is set as part of jwt config upon engine startup)",
)
Caching, Rate Limiting, Metrics & Session Variable Improvements (#376) * server: use a leaky bucket algorithm for bytes-per-second cache rate limiting * Use evalsha properly * Adds redis cache limit parameters to PoliciesConfig * Loads Leaky Bucket Script On Server Start * Adds more redis logging and moves cache update into lua script * reverts setex in lua and adds notes * Refactors cacheStore and adds max TTL and cache size limits * Filter session vars in cache key * WIP * parens * cache-clear-hander POC implementation * cache-clear-hander POC implementation * Pro projectId used as cache key * POC working! * prefixing query-response keys in redis * Add cacheClearer to RedisScripts * Partial implementation of cacheClearer from scripts record * updating tests * [automated] stylish-haskell commit * Adds query look with up with metrics script * Adds missing module and lua script from last commit * Changes redis script module structure to match cache clearing branch * minor change to lua script * cleaning up cache clearing * generalising JsonLog * [automated] stylish-haskell commit * Draft Cache Metrics Endpoint * Adds Cache Metrics Handler * Adds hook handler module * Missed HandlerHook module in last commit * glob * Fixes redis mget bug * Removes cache totals and changes dashes to colons in metric cache keys * Adds query param to clear clear endpoint for deleting specific keys * Adds query param to clear clear endpoint for deleting specific keys * Cache Metrics on query families rather then queries * Replace Set with nub * Base16 Redis Hashes * Query Family Redis Keys With Roles * response headers for cache keys * fixing bug in family key by excluding operation name; using hash for response header instead of entire key * Adds query family to redis cache keys and cache clear endpoint * Fixes queryfamily hash bug * Moves cache endpoints to /pro * Moved cache clear to POST * Refactors cache clear function * Fixes query family format bug * Adds query cache tests and optional --redis-url flag to python test suite * Adds session variable cache test * Update pro changelog * adding documentation for additional caching features * more docs * clearing up units of leaky bucket params * Adds comments to leaky bucket script * removes old todo * Fixes session variable filtering to work with new query rootfield * more advanced defaulting behaviour for bucket rate and capacity. * Updates Docs * Moves Role into QueryFamily hash * Use Aeson for Cache Clear endpoint response * Moves trace to bracket the leaky bucket script * Misc review tweaks * Adds sum type for cache clear query params * Hardcodes RegisReplyLog log level * Update docs/graphql/cloud/response-caching.rst Co-authored-by: Phil Freeman <phil@hasura.io> * new prose for rate limiting docs * [automated] stylish-haskell commit * make rootToSessVarPreds total * [automated] stylish-haskell commit * Fixes out of scope error * Renamed _acRedis to _acCacheStore Co-authored-by: Solomon Bothwell <ssbothwell@gmail.com> Co-authored-by: Lyndon Maydwell <lyndon@sordina.net> Co-authored-by: David Overton <david@hasura.io> Co-authored-by: Stylish Haskell Bot <stylish-haskell@users.noreply.github.com> Co-authored-by: Lyndon Maydwell <lyndon@hasura.io> GitOrigin-RevId: dda5c1a3f902967b3d78310f950541a55fabb1b0
2021-02-13 03:05:23 +03:00
parser.addoption(
"--redis-url",
metavar="REDIS_URL",
help="redis url for cache server",
default=False
)
parser.addoption(
"--backend",
help="run integration tests using a particular backend",
default="postgres"
)
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
parser.addoption(
"--pro-tests",
action="store_true",
default=False,
help="Flag to specify if the pro tests are to be run"
)
parser.addoption(
"--test-read-only-source",
action="store_true",
default=False,
required=False,
help="Run testcases with a read-only database source"
)
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
#By default,
#1) Set test grouping to by class (--dist=loadfile)
#2) Set default parallelism to one
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
def pytest_cmdline_preparse(config, args):
worker = os.environ.get('PYTEST_XDIST_WORKER')
if 'xdist' in sys.modules and not worker: # pytest-xdist plugin
num = 1
args[:] = ['--dist=loadfile', f'-n{num}'] + args
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
def pytest_configure(config):
# Pytest has removed the global pytest.config
# As a solution we are going to store it in PytestConf.config
PytestConf.config = config
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
if is_help_option_present(config):
return
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
if is_master(config):
assert not config.getoption('--exitfirst'), 'The "--exitfirst"/"-x" option does not work with xdist.\nSee: https://github.com/pytest-dev/pytest-xdist/issues/54'
if not (config.getoption('--hge-bin') or config.getoption('--hge-urls')):
print("either --hge-bin or --hge-urls should be specified")
if config.getoption('--hge-bin') and config.getoption('--hge-urls'):
print("only one of --hge-bin or --hge-urls should be specified")
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
if not config.getoption('--pg-urls'):
print("pg-urls should be specified")
config.hge_url_list = config.getoption('--hge-urls')
Caching, Rate Limiting, Metrics & Session Variable Improvements (#376) * server: use a leaky bucket algorithm for bytes-per-second cache rate limiting * Use evalsha properly * Adds redis cache limit parameters to PoliciesConfig * Loads Leaky Bucket Script On Server Start * Adds more redis logging and moves cache update into lua script * reverts setex in lua and adds notes * Refactors cacheStore and adds max TTL and cache size limits * Filter session vars in cache key * WIP * parens * cache-clear-hander POC implementation * cache-clear-hander POC implementation * Pro projectId used as cache key * POC working! * prefixing query-response keys in redis * Add cacheClearer to RedisScripts * Partial implementation of cacheClearer from scripts record * updating tests * [automated] stylish-haskell commit * Adds query look with up with metrics script * Adds missing module and lua script from last commit * Changes redis script module structure to match cache clearing branch * minor change to lua script * cleaning up cache clearing * generalising JsonLog * [automated] stylish-haskell commit * Draft Cache Metrics Endpoint * Adds Cache Metrics Handler * Adds hook handler module * Missed HandlerHook module in last commit * glob * Fixes redis mget bug * Removes cache totals and changes dashes to colons in metric cache keys * Adds query param to clear clear endpoint for deleting specific keys * Adds query param to clear clear endpoint for deleting specific keys * Cache Metrics on query families rather then queries * Replace Set with nub * Base16 Redis Hashes * Query Family Redis Keys With Roles * response headers for cache keys * fixing bug in family key by excluding operation name; using hash for response header instead of entire key * Adds query family to redis cache keys and cache clear endpoint * Fixes queryfamily hash bug * Moves cache endpoints to /pro * Moved cache clear to POST * Refactors cache clear function * Fixes query family format bug * Adds query cache tests and optional --redis-url flag to python test suite * Adds session variable cache test * Update pro changelog * adding documentation for additional caching features * more docs * clearing up units of leaky bucket params * Adds comments to leaky bucket script * removes old todo * Fixes session variable filtering to work with new query rootfield * more advanced defaulting behaviour for bucket rate and capacity. * Updates Docs * Moves Role into QueryFamily hash * Use Aeson for Cache Clear endpoint response * Moves trace to bracket the leaky bucket script * Misc review tweaks * Adds sum type for cache clear query params * Hardcodes RegisReplyLog log level * Update docs/graphql/cloud/response-caching.rst Co-authored-by: Phil Freeman <phil@hasura.io> * new prose for rate limiting docs * [automated] stylish-haskell commit * make rootToSessVarPreds total * [automated] stylish-haskell commit * Fixes out of scope error * Renamed _acRedis to _acCacheStore Co-authored-by: Solomon Bothwell <ssbothwell@gmail.com> Co-authored-by: Lyndon Maydwell <lyndon@sordina.net> Co-authored-by: David Overton <david@hasura.io> Co-authored-by: Stylish Haskell Bot <stylish-haskell@users.noreply.github.com> Co-authored-by: Lyndon Maydwell <lyndon@hasura.io> GitOrigin-RevId: dda5c1a3f902967b3d78310f950541a55fabb1b0
2021-02-13 03:05:23 +03:00
config.pg_url_list = config.getoption('--pg-urls')
if not config.getoption('--hge-bin') and config.getoption('-n', default=None):
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
xdist_threads = config.getoption('-n')
assert config.getoption('--hge-bin') or xdist_threads <= len(config.hge_url_list), "Not enough hge_urls specified, Required " + str(xdist_threads) + ", got " + str(len(config.hge_url_list))
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
assert xdist_threads <= len(config.pg_url_list), "Not enough pg_urls specified, Required " + str(xdist_threads) + ", got " + str(len(config.pg_url_list))
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
@pytest.hookimpl()
def pytest_report_collectionfinish(config, startdir, items):
"""
Collect server upgrade tests to the given file
"""
tests_file = config.getoption('--collect-upgrade-tests-to-file')
tests = collections.OrderedDict()
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
if tests_file:
def is_upgrade_test(item):
# Check if allow_server_upgrade_tests marker are present
# skip_server_upgrade_tests marker is not present
return item.get_closest_marker('allow_server_upgrade_test') \
and not item.get_closest_marker('skip_server_upgrade_test')
with open(tests_file,'w') as f:
upgrade_items = filter(is_upgrade_test, items)
for item in upgrade_items:
# This test should be run separately,
# since its schema setup has function scope
if 'per_method_tests_db_state' in item.fixturenames:
tests[item.nodeid] = True
elif any([ (x in item.fixturenames)
for x in
[ 'per_class_tests_db_state',
'per_class_db_schema_for_mutation_tests'
]
]):
# For this test, schema setup has class scope
# We can run a class of these tests at a time
tests[item.parent.nodeid] = True
# Assume tests can only be run separately
else:
tests[item.nodeid] = True
for test in tests.keys():
f.write(test + '\n')
return ''
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
@pytest.hookimpl(optionalhook=True)
def pytest_configure_node(node):
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
if is_help_option_present(node.config):
return
if not node.config.getoption('--hge-bin'):
node.workerinput["hge-url"] = node.config.hge_url_list.pop()
node.workerinput["pg-url"] = node.config.pg_url_list.pop()
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
def run_on_current_backend(request: pytest.FixtureRequest):
current_backend = request.config.getoption('--backend')
# Currently, we default all tests to run on Postgres with or without a --backend flag.
# As our test suite develops, we may consider running backend-agnostic tests on all
# backends, unless a specific `--backend` flag is passed.
desired_backends = set(name for marker in request.node.iter_markers('backend') for name in marker.args) or set(['postgres'])
return current_backend in desired_backends
def per_backend_tests_fixture(request: pytest.FixtureRequest):
"""
This fixture ignores backend-specific tests unless the relevant --backend flag has been passed.
"""
if not run_on_current_backend(request):
desired_backends = set(name for marker in request.node.iter_markers('backend') for name in marker.args)
pytest.skip('Skipping test. This test can run on ' + ', '.join(desired_backends) + '.')
@pytest.fixture(scope='class', autouse=True)
def per_backend_test_class(request: pytest.FixtureRequest):
return per_backend_tests_fixture(request)
@pytest.fixture(scope='function', autouse=True)
def per_backend_test_function(request: pytest.FixtureRequest):
return per_backend_tests_fixture(request)
@pytest.fixture(scope='session')
def pg_version(owner_engine: sqlalchemy.engine.Engine) -> int:
return fixtures.postgres.version(owner_engine)
@pytest.fixture(scope='session')
def owner_engine(request: pytest.FixtureRequest, hge_bin: str):
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
if not hge_bin:
return sqlalchemy.engine.create_engine(request.config.workerinput["pg-url"]) # type: ignore
return fixtures.postgres.owner_engine(request)
@pytest.fixture(scope='session')
def runner_engine(owner_engine: sqlalchemy.engine.Engine):
return fixtures.postgres.runner_engine(owner_engine)
@pytest.fixture(scope='class')
def metadata_schema_url(
request: pytest.FixtureRequest,
owner_engine: sqlalchemy.engine.Engine,
runner_engine: sqlalchemy.engine.Engine,
hge_bin: Optional[str],
):
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
if not hge_bin:
return request.config.workerinput["pg-url"] # type: ignore
return fixtures.postgres.metadata_schema_url(request, owner_engine, runner_engine)
@pytest.fixture(scope='class')
def source_backend(
request: pytest.FixtureRequest,
owner_engine: sqlalchemy.engine.Engine,
runner_engine: sqlalchemy.engine.Engine,
hge_ctx_fixture: HGECtx,
hge_bin: Optional[str],
):
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
if not hge_bin:
yield None
return
yield from fixtures.postgres.source_backend(request, owner_engine, runner_engine, hge_ctx_fixture)
@pytest.fixture(scope='class')
def postgis(owner_engine: sqlalchemy.engine.Engine, source_backend: fixtures.postgres.Backend):
return fixtures.postgres.postgis(owner_engine, source_backend)
@pytest.fixture(scope='session')
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
def tls_ca_configuration(request: pytest.FixtureRequest, tmp_path_factory: pytest.TempPathFactory) -> Optional[fixtures.tls.TLSCAConfiguration]:
cert_file: Optional[str] = request.config.getoption('--tls-ca-cert') # type: ignore
key_file: Optional[str] = request.config.getoption('--tls-ca-key') # type: ignore
if cert_file and key_file:
tmp_path = tmp_path_factory.mktemp('tls')
return fixtures.tls.TLSCAConfiguration(cert_file, key_file, tmp_path)
else:
return None
# TODO: remove once parallelization work is completed
@pytest.fixture(scope='class', autouse=True)
def hge_skip(request: pytest.FixtureRequest, hge_server: Optional[str], hge_fixture_env: dict[str, str]):
# Let `hge_server` manage this stuff.
if hge_server:
return
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
# Ensure that the correct environment variables have been set for the given test.
hge_marker_env: dict[str, str] = {marker.args[0]: marker.args[1] for marker in request.node.iter_markers('hge_env')}
hge_env = {**hge_marker_env, **hge_fixture_env}
incorrect_env = {name: value for name, value in hge_env.items() if os.getenv(name) != value}
if len(incorrect_env) > 0:
pytest.skip(
'This test expects the following environment variables: '
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
+ ', '.join([f'{name!r} = {value!r} (not {os.getenv(name)!r})' for name, value in incorrect_env.items()]))
# Make sure that if there's a webhook, the test expects that there's a webhook
if 'HASURA_GRAPHQL_AUTH_HOOK' in os.environ:
if 'HASURA_GRAPHQL_AUTH_HOOK' not in hge_fixture_env:
pytest.skip('HGE expects a running webhook, but this test does not provide one.')
if os.environ['HASURA_GRAPHQL_AUTH_HOOK'] != hge_fixture_env['HASURA_GRAPHQL_AUTH_HOOK']:
pytest.skip(f'HGE expects a running webhook at {os.environ["HASURA_GRAPHQL_AUTH_HOOK"]}, but this test provides one at {hge_fixture_env["HASURA_GRAPHQL_AUTH_HOOK"]}.')
@pytest.fixture(scope='session')
def hge_bin(request: pytest.FixtureRequest) -> Optional[str]:
return request.config.getoption('--hge-bin') # type: ignore
@pytest.fixture(scope='class')
def hge_port() -> int:
return fixtures.hge.hge_port()
@pytest.fixture(scope='class')
def hge_url(request: pytest.FixtureRequest, hge_bin: Optional[str], hge_port: int) -> str:
if hge_bin:
return f'http://localhost:{hge_port}'
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
else:
return request.config.workerinput['hge-url'] # type: ignore
# A hack to inject environment variables from other fixtures into HGE.
# All of this is because we cannot cleanly express dependencies between
# fixtures of the form "this loads before that, IF that is loaded".
#
# This is marked as "early" so the `fixture-order` plugin ensures that it is
# loaded _before_ `hge_server`. Similarly, any fixture using it must be at
# the same scope level and also marked as "early", to ensure that it is
# mutated before `hge_server` uses the data.
#
# In short, we use `early` to ensure that writes happen before reads.
@pytest.fixture(scope='class')
@pytest.mark.early
def hge_fixture_env() -> dict[str, str]:
return {}
@pytest.fixture(scope='class')
def hge_key(
request: pytest.FixtureRequest,
hge_bin: Optional[str],
) -> Optional[str]:
marker = request.node.get_closest_marker('admin_secret')
if hge_bin:
# If the test requests an admin secret, generate one.
return str(uuid.uuid4()) if marker else None
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
else:
# If the environment variable is set, use it.
# This will be used in the event that we start the server outside the test harness.
# We skip the test if it explicitly requires that we have no admin secret.
anti_marker = request.node.get_closest_marker('no_admin_secret')
env_key = os.environ.get('HASURA_GRAPHQL_ADMIN_SECRET')
if anti_marker and env_key:
pytest.skip('This test requires that the admin secret is not set.')
return env_key
@pytest.fixture(scope='class')
def hge_server(
request: pytest.FixtureRequest,
hge_bin: Optional[str],
hge_port: int,
hge_url: str,
hge_key: Optional[str],
hge_fixture_env: dict[str, str],
metadata_schema_url: str,
) -> Optional[str]:
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
if not hge_bin:
return None
return fixtures.hge.hge_server(request, hge_bin, hge_port, hge_url, hge_key, hge_fixture_env, metadata_schema_url)
@pytest.fixture(scope='class')
def enabled_apis(request: pytest.FixtureRequest, hge_bin: Optional[str]) -> Optional[set[str]]:
if hge_bin:
hge_marker_env: dict[str, str] = {marker.args[0]: marker.args[1] for marker in request.node.iter_markers('hge_env') if marker.args[1] is not None}
enabled_apis_str = hge_marker_env.get('HASURA_GRAPHQL_ENABLED_APIS')
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
else:
enabled_apis_str = os.environ.get('HASURA_GRAPHQL_ENABLED_APIS')
if not enabled_apis_str:
return None
return set(enabled_apis_str.split(','))
@pytest.fixture(scope='class')
def hge_ctx_fixture(
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
request: pytest.FixtureRequest,
hge_url: str,
hge_bin: Optional[str],
metadata_schema_url: str,
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
hge_key: Optional[str],
enabled_apis: Optional[set[str]],
):
# This madness allows us to figure out whether there is a webhook running.
# We need this information because we dynamically decide how we run queries according to the authentication method.
# This is probably terrible, but refactoring that logic would require rewriting every test.
webhook: Optional[HGECtxWebhook] = None
if webhook_server.__name__ in request.fixturenames:
webhook = request.getfixturevalue(webhook_server.__name__)
elif 'query_echo_webhook' in request.fixturenames: # in test_webhook_request_context.py
webhook = HGECtxWebhook(tls_trust=None)
hge_ctx = HGECtx(
hge_url=hge_url,
metadata_schema_url=metadata_schema_url,
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
hge_key=hge_key,
webhook=webhook,
enabled_apis=enabled_apis,
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
clear_dbs=not hge_bin,
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
config=request.config,
)
yield hge_ctx
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
2018-09-18 09:21:57 +03:00
hge_ctx.teardown()
time.sleep(1) # TODO why do we sleep here?
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
# tie everything together
@pytest.fixture(scope='class')
def hge_ctx(
hge_ctx_fixture: HGECtx,
hge_server: Optional[str],
source_backend: Optional[fixtures.postgres.Backend],
):
return hge_ctx_fixture
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
@pytest.fixture(scope='class')
@pytest.mark.early
def evts_webhook(hge_fixture_env: dict[str, str]):
server = EvtsWebhookServer(extract_server_address_from('EVENT_WEBHOOK_HANDLER'))
thread = threading.Thread(target=server.serve_forever)
thread.start()
print(f'{evts_webhook.__name__} server started on {server.url}')
hge_fixture_env['EVENT_WEBHOOK_HANDLER'] = server.url
yield server
server.shutdown()
server.server_close()
thread.join()
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
@pytest.fixture(scope='class')
@pytest.mark.early
def actions_fixture(hge_url: str, hge_key: Optional[str], hge_fixture_env: dict[str, str]):
# Start actions' webhook server
server = ActionsWebhookServer(hge_url, hge_key, server_address=extract_server_address_from('ACTION_WEBHOOK_HANDLER'))
thread = threading.Thread(target=server.serve_forever)
thread.start()
print(f'{actions_fixture.__name__} server started on {server.url}')
hge_fixture_env['ACTION_WEBHOOK_HANDLER'] = server.url
yield server
server.shutdown()
server.server_close()
thread.join()
allow custom mutations through actions (#3042) * basic doc for actions * custom_types, sync and async actions * switch to graphql-parser-hs on github * update docs * metadata import/export * webhook calls are now supported * relationships in sync actions * initialise.sql is now in sync with the migration file * fix metadata tests * allow specifying arguments of actions * fix blacklist check on check_build_worthiness job * track custom_types and actions related tables * handlers are now triggered on async actions * default to pgjson unless a field is involved in relationships, for generating definition list * use 'true' for action filter for non admin role * fix create_action_permission sql query * drop permissions when dropping an action * add a hdb_role view (and relationships) to fetch all roles in the system * rename 'webhook' key in action definition to 'handler' * allow templating actions wehook URLs with env vars * add 'update_action' /v1/query type * allow forwarding client headers by setting `forward_client_headers` in action definition * add 'headers' configuration in action definition * handle webhook error response based on status codes * support array relationships for custom types * implement single row mutation, see https://github.com/hasura/graphql-engine/issues/3731 * single row mutation: rename 'pk_columns' -> 'columns' and no-op refactor * use top level primary key inputs for delete_by_pk & account select permissions for single row mutations * use only REST semantics to resolve the webhook response * use 'pk_columns' instead of 'columns' for update_by_pk input * add python basic tests for single row mutations * add action context (name) in webhook payload * Async action response is accessible for non admin roles only if the request session vars equals to action's * clean nulls, empty arrays for actions, custom types in export metadata * async action mutation returns only the UUID of the action * unit tests for URL template parser * Basic sync actions python tests * fix output in async query & add async tests * add admin secret header in async actions python test * document async action architecture in Resolve/Action.hs file * support actions returning array of objects * tests for list type response actions * update docs with actions and custom types metadata API reference * update actions python tests as per #f8e1330 Co-authored-by: Tirumarai Selvan <tirumarai.selvan@gmail.com> Co-authored-by: Aravind Shankar <face11301@gmail.com> Co-authored-by: Rakesh Emmadi <12475069+rakeshkky@users.noreply.github.com>
2020-02-13 20:38:23 +03:00
use_action_fixtures = pytest.mark.usefixtures(
'actions_fixture',
'per_class_db_schema_for_mutation_tests',
'per_method_db_data_for_mutation_tests'
)
@pytest.fixture(scope='class')
@pytest.mark.early
def auth_hook(hge_fixture_env: dict[str, str]):
server = auth_webhook_server.create_server(extract_server_address_from('HASURA_GRAPHQL_AUTH_HOOK'))
thread = threading.Thread(target = server.serve_forever)
thread.start()
print(f'{auth_hook.__name__} server started on {server.url}')
hge_fixture_env['HASURA_GRAPHQL_AUTH_HOOK'] = server.url + '/auth'
ports.wait_for_port(server.server_port)
yield server
auth_webhook_server.stop_server(server)
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
@pytest.fixture(scope='class')
@pytest.mark.early
def webhook_server(
request: pytest.FixtureRequest,
hge_bin: Optional[str],
hge_fixture_env: dict[str,str],
tls_ca_configuration: Optional[fixtures.tls.TLSCAConfiguration],
):
server_address = extract_server_address_from('HASURA_GRAPHQL_AUTH_HOOK')
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
scheme = None
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
if not hge_bin:
scheme = str(urllib.parse.urlparse(os.getenv('HASURA_GRAPHQL_AUTH_HOOK')).scheme)
if tls_ca_configuration:
if scheme is not None and scheme != 'https':
pytest.skip(f'Cannot run the remote schema server with TLS; HGE is configured to talk to it over "{scheme}".')
if request.node.get_closest_marker('no_tls_webhook_server') is not None:
pytest.skip('Only running this test with TLS disabled; skipping the version with TLS enabled.')
server = http.server.HTTPServer(server_address, webhook.Handler)
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
insecure = request.node.get_closest_marker('tls_insecure_certificate') is not None
tls_trust = fixtures.tls.TLSTrust.INSECURE if insecure else fixtures.tls.TLSTrust.SECURE
tls_ca_configuration.configure(server, tls_trust)
else:
if scheme is not None and scheme != 'http':
pytest.skip(f'Cannot run the remote schema server without TLS; HGE is configured to talk to it over "{scheme}".')
if request.node.get_closest_marker('tls_webhook_server') is not None:
pytest.skip('Only running this test with TLS enabled; skipping the version with TLS disabled.')
server = http.server.HTTPServer(server_address, webhook.Handler)
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
tls_trust = None
thread = threading.Thread(target=server.serve_forever)
thread.start()
request.addfinalizer(server.shutdown)
scheme = 'https' if tls_trust else 'http'
# We must use 'localhost' and not `server.server_address[0]`
# because when using TLS, we need a domain name, not an IP address.
host = 'localhost'
port = server.server_address[1]
hge_fixture_env['HASURA_GRAPHQL_AUTH_HOOK'] = f'{scheme}://{host}:{port}/'
ports.wait_for_port(port)
return HGECtxWebhook(tls_trust=tls_trust)
@pytest.fixture(scope='class')
def pro_tests_fixtures(hge_ctx):
if not hge_ctx.pro_tests:
pytest.skip('These tests are meant to be run with --pro-tests set')
@pytest.fixture(scope='class')
@pytest.mark.early
def scheduled_triggers_evts_webhook(hge_fixture_env: dict[str, str]):
server = EvtsWebhookServer(extract_server_address_from('SCHEDULED_TRIGGERS_WEBHOOK_DOMAIN'))
thread = threading.Thread(target=server.serve_forever)
thread.start()
print(f'{scheduled_triggers_evts_webhook.__name__} server started on {server.url}')
hge_fixture_env['SCHEDULED_TRIGGERS_WEBHOOK_DOMAIN'] = server.url
yield server
server.shutdown()
server.server_close()
thread.join()
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
# This will use TLS on CI, and won't when run locally. Unfortunately,
# parameterization doesn't really work (see test_webhook.py and test_tests.py
# for details), so we shall avoid it in favor of just testing what we can.
@pytest.fixture(scope='class')
@pytest.mark.early
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
def gql_server(
request: pytest.FixtureRequest,
hge_bin: Optional[str],
hge_url: str,
hge_fixture_env: dict[str,str],
tls_ca_configuration: Optional[fixtures.tls.TLSCAConfiguration],
):
server_address = extract_server_address_from('REMOTE_SCHEMAS_WEBHOOK_DOMAIN')
scheme = None
# TODO: remove once parallelization work is completed
# `hge_bin` will no longer be optional
if not hge_bin:
scheme = str(urllib.parse.urlparse(os.getenv('REMOTE_SCHEMAS_WEBHOOK_DOMAIN')).scheme)
if socket.socket(socket.AF_INET, socket.SOCK_STREAM).connect_ex(server_address) == 0:
# The server is already running. This might be the case if we're running the server upgrade/downgrade tests.
# In this case, skip starting it and rely on the external service.
url = f'http://{server_address[0]}:{server_address[1]}'
hge_fixture_env['REMOTE_SCHEMAS_WEBHOOK_DOMAIN'] = url
# Create an object with a field named "url", set to `url`.
return collections.namedtuple('ExternalGraphQLServer', ['url'])(url)
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
if tls_ca_configuration:
if scheme is not None and scheme != 'https':
pytest.skip(f'Cannot run the remote schema server with TLS; HGE is configured to talk to it over "{scheme}".')
else:
if scheme is not None and scheme != 'http':
pytest.skip(f'Cannot run the remote schema server without TLS; HGE is configured to talk to it over "{scheme}".')
hge_urls: list[str] = request.config.getoption('--hge-urls') or [hge_url] # type: ignore
server = HGECtxGQLServer(
server_address=server_address,
tls_ca_configuration=tls_ca_configuration,
hge_urls=hge_urls,
)
server.start_server()
print(f'{gql_server.__name__} server started on {server.url}')
hge_fixture_env['REMOTE_SCHEMAS_WEBHOOK_DOMAIN'] = server.url
request.addfinalizer(server.stop_server)
return server
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
@pytest.fixture(scope='class')
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
def ws_client(hge_ctx):
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
"""
This fixture provides an Apollo GraphQL websockets client
"""
client = GQLWsClient(hge_ctx, '/v1/graphql')
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
time.sleep(0.1)
yield client
client.teardown()
@pytest.fixture(scope='class')
server/tests-py: Start webhook.py inside the test harness. We use a helper service to start a webhook-based authentication service for some tests. This moves the initialization of the service out of _test-server.sh_ and into the Python test harness, as a fixture. In order to do this, I had to make a few changes. The main deviation is that we no longer run _all_ tests against an HGE with this authentication service, just a few (those in _test_webhook.py_). Because this reduced coverage, I have added some more tests there, which actually cover some areas not exacerbated elsewhere (mainly trying to use webhook credentials to talk to an admin-only endpoint). The webhook service can run both with and without TLS, and decide whether it's necessary to skip one of these based on the arguments passed and how HGE is started, according to the following logic: * If a TLS CA certificate is passed in, it will run with TLS, otherwise it will skip it. * If HGE was started externally and a TLS certificate is provided, it will skip running without TLS, as it will assume that HGE was configured to talk to a webhook over HTTPS. * Some tests should only be run with TLS; this is marked with a `tls_webhook_server` marker. * Some tests should only be run _without_ TLS; this is marked with a `no_tls_webhook_server` marker. The actual parameterization of the webhook service configuration is done through test subclasses, because normal pytest parameterization doesn't work with the `hge_fixture_env` hack that we use. Because `hge_fixture_env` is not a sanctioned way of conveying data between fixtures (and, unfortunately, there isn't a sanctioned way of doing this when the fixtures in question may not know about each other directly), parameterizing the `webhook_server` fixture doesn't actually parameterize `hge_server` properly. Subclassing forces this to work correctly. The certificate generation is moved to a Python fixture, so that we don't have to revoke the CA certificate for _test_webhook_insecure.py_; we can just generate a bogus certificate instead. The CA certificate is still generated in the _test-server.sh_ script, as it needs to be installed into the OS certificate store. Interestingly, the CA certificate installation wasn't actually working, because the certificates were written to the wrong location. This didn't cause any failures, as we weren't actually testing this behavior. This is now fixed with the other changes. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6363 GitOrigin-RevId: 0f277d374daa64f657257ed2a4c2057c74b911db
2022-10-20 21:58:36 +03:00
def ws_client_graphql_ws(hge_ctx):
"""
This fixture provides an GraphQL-WS client
"""
client = GraphQLWSClient(hge_ctx, '/v1/graphql')
time.sleep(0.1)
yield client
client.teardown()
def extract_server_address_from(env_var: str) -> tuple[str, int]:
"""
Extracts a server address (a pair of host and port) from the given environment variable.
If the environment variable doesn't exist, returns `('localhost', 0)`.
"""
value = os.getenv(env_var)
if not value:
return ('localhost', 0)
url = urllib.parse.urlparse(value)
if not url.hostname:
raise Exception(f'Invalid host in {env_var}.')
if not url.port:
raise Exception(f'Invalid port in {env_var}.')
print(f'Found {env_var} = {value!r}, so {inspect.stack()[1].function} server will start on {url.geturl()}')
return (url.hostname, url.port)
@pytest.fixture(scope='class')
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
def per_class_tests_db_state(request, hge_ctx):
"""
Set up the database state for select queries.
Has a class level scope, since select queries does not change database state
Expects either `dir()` method which provides the directory
with `setup.yaml` and `teardown.yaml` files
Or class variables `setup_files` and `teardown_files` that provides
the list of setup and teardown files respectively.
By default, for a postgres backend the setup and teardown is done via
the `/v1/query` endpoint, to setup using the `/v1/metadata` (metadata setup)
and `/v2/query` (DB setup), set the `setup_metadata_api_version` to "v2"
"""
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
yield from db_state_context(request, hge_ctx)
@pytest.fixture(scope='function')
def per_method_tests_db_state(request, hge_ctx):
"""
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
This fixture sets up the database state for metadata operations
Has a function level scope, since metadata operations may change both the schema and data
Class method/variable requirements are similar to that of per_class_tests_db_state fixture
"""
yield from db_state_context(request, hge_ctx)
@pytest.fixture(scope='class')
def per_class_db_schema_for_mutation_tests(request, hge_ctx):
"""
This fixture sets up the database schema for mutations.
It has a class level scope, since mutations does not change schema.
Expects either `dir()` class method which provides the directory with `schema_setup.yaml` and `schema_teardown.yaml` files,
or variables `schema_setup_files` and `schema_teardown_files`
that provides the list of setup and teardown files respectively
"""
# setting the default metadata API version to v1
setup_metadata_api_version = getattr(request.cls, 'setup_metadata_api_version',"v1")
(setup, teardown, schema_setup, schema_teardown, pre_setup, post_teardown) = [
hge_ctx.backend_suffix(filename) + ".yaml"
for filename in ['setup', 'teardown', 'schema_setup', 'schema_teardown', 'pre_setup', 'post_teardown']
]
if hge_ctx.is_default_backend:
if setup_metadata_api_version == "v1":
db_context = db_context_with_schema_common(
request, hge_ctx,
'schema_setup_files', 'schema_setup.yaml',
'schema_teardown_files', 'schema_teardown.yaml',
)
else:
db_context = db_context_with_schema_common_new(
request, hge_ctx,
'schema_setup_files', setup,
'schema_teardown_files', teardown,
schema_setup, schema_teardown,
pre_setup, post_teardown,
)
else:
db_context = db_context_with_schema_common_new(
request, hge_ctx,
'schema_setup_files', setup,
'schema_teardown_files', teardown,
schema_setup, schema_teardown,
pre_setup, post_teardown,
)
yield from db_context
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
@pytest.fixture(scope='function')
def per_method_db_data_for_mutation_tests(request, hge_ctx, per_class_db_schema_for_mutation_tests):
"""
This fixture sets up the data for mutations.
Has a function level scope, since mutations may change data.
Having just the setup file(s), or the teardown file(s) is allowed.
Expects either `dir()` class method which provides the directory with `values_setup.yaml` and / or `values_teardown.yaml` files.
The class may provide `values_setup_files` variables which contains the list of data setup files,
Or the `values_teardown_files` variable which provides the list of data teardown files.
"""
# Non-default (Postgres) backend tests expect separate setup and schema_setup
# files for v1/metadata and v2/query requests, respectively.
(values_setup, values_teardown) = [
hge_ctx.backend_suffix(filename) + ".yaml"
for filename in ['values_setup', 'values_teardown']
]
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
yield from db_context_common(
request, hge_ctx,
'values_setup_files', values_setup,
'values_teardown_files', values_teardown,
skip_setup=False, skip_teardown=False
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
)
def db_state_context(request, hge_ctx):
# Non-default (Postgres) backend tests expect separate setup and schema_setup
# files for v1/metadata and v2/query requests, respectively.
(setup, teardown, schema_setup, schema_teardown, pre_setup, post_teardown) = [
hge_ctx.backend_suffix(filename) + ".yaml"
for filename in ['setup', 'teardown', 'schema_setup', 'schema_teardown', 'pre_setup', 'post_teardown']
]
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
# setting the default metadata API version to v1
setup_metadata_api_version = getattr(request.cls, 'setup_metadata_api_version',"v1")
if hge_ctx.is_default_backend:
if setup_metadata_api_version == "v1":
# setup the metadata and DB schema using the `/v1/query` endpoint
db_context = db_context_with_schema_common(
request, hge_ctx,
'setup_files', 'setup.yaml',
'teardown_files', 'teardown.yaml',
)
elif setup_metadata_api_version == "v2":
# setup the metadata using the "/v1/metadata" and the DB schema using the `/v2/query` endpoints
db_context = db_context_with_schema_common_new(
request, hge_ctx,
'setup_files', setup,
'teardown_files', teardown,
schema_setup, schema_teardown,
pre_setup, post_teardown,
)
else:
raise NotImplementedError('Invalid API version.')
else:
# setup the metadata using the "/v1/metadata" and the DB schema using the `/v2/query` endpoints
db_context = db_context_with_schema_common_new(
request, hge_ctx,
'setup_files', setup,
'teardown_files', teardown,
schema_setup, schema_teardown,
pre_setup, post_teardown,
)
yield from db_context
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
def db_context_with_schema_common(
request, hge_ctx,
setup_files_attr, setup_default_file,
teardown_files_attr, teardown_default_file,
):
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
(skip_setup, skip_teardown) = [
request.config.getoption('--' + x)
for x in ['skip-schema-setup', 'skip-schema-teardown']
]
yield from db_context_common(
request, hge_ctx,
setup_files_attr, setup_default_file,
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
teardown_files_attr, teardown_default_file,
skip_setup, skip_teardown
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
)
def db_context_with_schema_common_new(
request, hge_ctx,
setup_files_attr, setup_default_file,
teardown_files_attr, teardown_default_file,
setup_sql_file, teardown_sql_file,
pre_setup_file, post_teardown_file,
):
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
(skip_setup, skip_teardown) = [
request.config.getoption('--' + x)
for x in ['skip-schema-setup', 'skip-schema-teardown']
]
yield from db_context_common_new(
request, hge_ctx,
setup_files_attr, setup_default_file, setup_sql_file,
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
teardown_files_attr, teardown_default_file, teardown_sql_file,
pre_setup_file, post_teardown_file,
skip_setup, skip_teardown
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
)
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
def db_context_common(
request, hge_ctx,
setup_files_attr, setup_default_file,
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
teardown_files_attr, teardown_default_file,
skip_setup=True, skip_teardown=True
):
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
def get_files(attr, default_file):
files = getattr(request.cls, attr, None)
if not files:
files = os.path.join(request.cls.dir(), default_file)
return files
setup = get_files(setup_files_attr, setup_default_file)
teardown = get_files(teardown_files_attr, teardown_default_file)
if hge_ctx.is_default_backend:
yield from setup_and_teardown_v1q(
request, hge_ctx,
setup, teardown,
skip_setup, skip_teardown
)
else:
yield from setup_and_teardown_v2q(
request, hge_ctx,
setup, teardown,
skip_setup, skip_teardown
)
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
def db_context_common_new(
request, hge_ctx,
setup_files_attr, setup_default_file, setup_default_sql_file,
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
teardown_files_attr, teardown_default_file, teardown_default_sql_file,
pre_setup_file, post_teardown_file,
skip_setup=True, skip_teardown=True
):
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
def get_files(attr, default_file):
files = getattr(request.cls, attr, None)
if not files:
files = os.path.join(request.cls.dir(), default_file)
return files
setup = get_files(setup_files_attr, setup_default_file)
teardown = get_files(teardown_files_attr, teardown_default_file)
setup_default_sql_file = os.path.join(request.cls.dir(), setup_default_sql_file)
teardown_default_sql_file = os.path.join(request.cls.dir(), teardown_default_sql_file)
pre_setup_default_file = os.path.join(request.cls.dir(), pre_setup_file)
post_teardown_default_file = os.path.join(request.cls.dir(), post_teardown_file)
yield from setup_and_teardown(
request, hge_ctx,
setup, teardown,
setup_default_sql_file, teardown_default_sql_file,
pre_setup_default_file, post_teardown_default_file,
skip_setup, skip_teardown
)
def setup_and_teardown_v1q(
request, hge_ctx,
setup_files, teardown_files,
skip_setup=False, skip_teardown=False
):
def v1q_f(filepath):
if os.path.isfile(filepath):
return hge_ctx.v1q_f(filepath)
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
if not skip_setup:
run_on_elem_or_list(v1q_f, setup_files)
yield
# Teardown anyway if any of the tests have failed
if request.session.testsfailed > 0 or not skip_teardown:
run_on_elem_or_list(v1q_f, teardown_files)
def setup_and_teardown_v2q(
request, hge_ctx,
setup_files, teardown_files,
skip_setup=False, skip_teardown=False
):
def v2q_f(filepath):
if os.path.isfile(filepath):
return hge_ctx.v2q_f(filepath)
if not skip_setup:
run_on_elem_or_list(v2q_f, setup_files)
yield
# Teardown anyway if any of the tests have failed
if request.session.testsfailed > 0 or not skip_teardown:
run_on_elem_or_list(v2q_f, teardown_files)
def setup_and_teardown(
request, hge_ctx,
setup_files, teardown_files,
sql_schema_setup_file, sql_schema_teardown_file,
pre_setup_file, post_teardown_file,
skip_setup=False, skip_teardown=False
):
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
def v2q_f(f):
if os.path.isfile(f):
try:
hge_ctx.v2q_f(f)
except AssertionError:
try:
run_on_elem_or_list(pre_post_metadataq_f, post_teardown_file)
except:
pass
raise
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
def metadataq_f(f):
if os.path.isfile(f):
try:
hge_ctx.v1metadataq_f(f)
except AssertionError:
try:
# drop the sql setup, if the metadata calls fail
run_on_elem_or_list(v2q_f, sql_schema_teardown_file)
run_on_elem_or_list(pre_post_metadataq_f, post_teardown_file)
except:
pass
raise
def pre_post_metadataq_f(f):
if os.path.isfile(f):
hge_ctx.v1metadataq_f(f)
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
if not skip_setup:
run_on_elem_or_list(pre_post_metadataq_f, pre_setup_file)
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
run_on_elem_or_list(v2q_f, sql_schema_setup_file)
run_on_elem_or_list(metadataq_f, setup_files)
yield
# Teardown anyway if any of the tests have failed
if request.session.testsfailed > 0 or not skip_teardown:
run_on_elem_or_list(metadataq_f, teardown_files)
run_on_elem_or_list(v2q_f, sql_schema_teardown_file)
run_on_elem_or_list(pre_post_metadataq_f, post_teardown_file)
[Preview] Inherited roles for postgres read queries fixes #3868 docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de` Note: To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`. Introduction ------------ This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`. How are select permissions of different roles are combined? ------------------------------------------------------------ A select permission includes 5 things: 1. Columns accessible to the role 2. Row selection filter 3. Limit 4. Allow aggregation 5. Scalar computed fields accessible to the role Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`. Let's say the following GraphQL query is queried with the `combined_roles` role. ```graphql query { employees { address phone } } ``` This will translate to the following SQL query: ```sql select (case when (P1 or P2) then address else null end) as address, (case when P2 then phone else null end) as phone from employee where (P1 or P2) ``` The other parameters of the select permission will be combined in the following manner: 1. Limit - Minimum of the limits will be the limit of the inherited role 2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation 3. Scalar computed fields - same as table column fields, as in the above example APIs for inherited roles: ---------------------- 1. `add_inherited_role` `add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments `role_name`: the name of the inherited role to be added (String) `role_set`: list of roles that need to be combined (Array of Strings) Example: ```json { "type": "add_inherited_role", "args": { "role_name":"combined_user", "role_set":[ "user", "user1" ] } } ``` After adding the inherited role, the inherited role can be used like single roles like earlier Note: An inherited role can only be created with non-inherited/singular roles. 2. `drop_inherited_role` The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument: `role_name`: name of the inherited role to be dropped Example: ```json { "type": "drop_inherited_role", "args": { "role_name":"combined_user" } } ``` Metadata --------- The derived roles metadata will be included under the `experimental_features` key while exporting the metadata. ```json { "experimental_features": { "derived_roles": [ { "role_name": "manager_is_employee_too", "role_set": [ "employee", "manager" ] } ] } } ``` Scope ------ Only postgres queries and subscriptions are supported in this PR. Important points: ----------------- 1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done. TODOs ------- - [ ] Tests - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?) - [ ] Introspection test with a inherited role (nullability changes in a inherited role) - [ ] Docs - [ ] Changelog Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com> GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 14:14:13 +03:00
run default tests in test_server_upgrade (#3718) * run basic tests after upgrade * terminate before specifying file in pytest cmd * Move fixture definitions out of test classes Previously we had abstract classes with the fixtures defined in them. The test classes then inherits these super classes. This is creating inheritence problems, especially when you want to just inherit the tests in class, but not the fixtures. We have now moved all those fixture definitions outside of the class (in conftest.py). These fixtures are now used by the test classes when and where they are required. * Run pytests on server upgrade Server upgrade tests are run by 1) Run pytest with schema/metadata setup but do not do schema/metadata teardown 2) Upgrade the server 3) Run pytest using the above schema and teardown at the end of the tests 4) Cleanup hasura metadata and start again with next set of tests We have added options --skip-schema-setup and --skip-schema-teardown to help running server upgrade tests. While running the tests, we noticed that error codes and messages for some of the tests have changed. So we have added another option to pytest `--avoid-error-message-checks`. If this flag is set, and if comparing expected and response message fails, and if the expected response has an error message, Pytest will throw warnings instead of an error. * Use marks to specify server-upgrade tests Not all tests can be run as serve upgrade tests, particularly those which themselves change the schema. We introduce two pytest markers. Marker allow_server_upgrade_test will add the test into the list of server upgrade tests that can be run. skip_server_upgrade_test removes it from the list. With this we have added tests for queries, mutations, and selected event trigger and remote schema tests to the list of server upgrade tests. * Remove components not needed anymore * Install curl * Fix error in query validation * Fix error in test_v1_queries.py * install procps for server upgrade tests * Use postgres image which has postgis installed * set pager off with psql * quote the bash variable WORKTREE_DIR Co-authored-by: nizar-m <19857260+nizar-m@users.noreply.github.com> Co-authored-by: Vamshi Surabhi <0x777@users.noreply.github.com>
2020-02-13 12:14:02 +03:00
def run_on_elem_or_list(f, x):
if isinstance(x, str):
return [f(x)]
elif isinstance(x, list):
return [f(e) for e in x]
def is_help_option_present(config):
return any([
config.getoption(x)
for x in ['--fixtures','--help', '--collect-only']
])
run graphql tests on both http and websocket; add parallelism (close #1868) (#1921) Examples 1) ` pytest --hge-urls "http://127.0.0.1:8080" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" -vv ` 2) `pytest --hge-urls "http://127.0.0.1:8080" "http://127.0.0.1:8081" --pg-urls "postgresql://admin@127.0.0.1:5432/hge_tests" "postgresql://admin@127.0.0.1:5432/hge_tests2" -vv ` ### Solution and Design <!-- How is this issue solved/fixed? What is the design? --> <!-- It's better if we elaborate --> #### Reducing execution time of tests - The Schema setup and teardown, which were earlier done per test method, usually takes around 1 sec. - For mutations, the model has now been changed to only do schema setup and teardown once per test class. - A data setup and teardown will be done once per test instead (usually takes ~10ms). - For the test class to get this behaviour, one can can extend the class `DefaultTestMutations`. - The function `dir()` should be define which returns the location of the configuration folder. - Inside the configuration folder, there should be - Files `<conf_dir>/schema_setup.yaml` and `<conf_dir>/schema_teardown.yaml`, which has the metadata query executed during schema setup and teardown respectively - Files named `<conf_dir>/values_setup.yaml` and `<conf_dir>/values_teardown.yaml`. These files are executed to setup and remove data from the tables respectively. #### Running Graphql queries on both http and websockets - Each GraphQL query/mutation is run on the both HTTP and websocket protocols - Pytests test parameterisation is used to achieve this - The errors over websockets are slightly different from that on HTTP - The code takes care of converting the errors in HTTP to errors in websockets #### Parallel executation of tests. - The plugin pytest-xdist helps in running tests on parallel workers. - We are using this plugin to group tests by file and run on different workers. - Parallel test worker processes operate on separate postgres databases(and separate graphql-engines connected to these databases). Thus tests on one worker will not affect the tests on the other worker. - With two workers, this decreases execution times by half, as the tests on event triggers usually takes a long time, but does not consume much CPU.
2019-04-08 10:22:38 +03:00
def is_master(config):
"""True if the code running the given pytest.config object is running in a xdist master
node or not running xdist at all.
"""
return not hasattr(config, 'workerinput')