This is a follow-up to #1959.
Today, I spent a while in review figuring out that a harmless PR change didn't do anything,
because it was moving from a `runLazy...` to something without the `Lazy`. So let's get
that source of confusion removed.
This should be a bit easier to review commit by commit, since some of the functions had
confusing names. (E.g. there was a misnamed `Migrate.Internal.runTx` before.)
The change should be a no-op.
https://github.com/hasura/graphql-engine-mono/pull/2335
GitOrigin-RevId: 0f284c4c0f814482d7827e7732a6d49e7735b302
>
### Description
>
This PR is an incremental work towards [enabling insert mutations on MSSQL](https://github.com/hasura/graphql-engine-mono/pull/1974). In this PR, we generate insert mutation schema parser for MSSQL backend.
### Changelog
- [ ] `CHANGELOG.md` is updated with user-facing content relevant to this PR. If no changelog is required, then add the `no-changelog-required` label.
### Affected components
- [x] Server
https://github.com/hasura/graphql-engine-mono/pull/2141
GitOrigin-RevId: 8595008dece35f7fded9c52e134de8b97b64f53f
When adding object relationships, we set the nullability of the generated GraphQL field based on whether the database backend enforces that the referenced data always exists. For manual relationships (corresponding to `manual_configuration`), the database backend is unaware of any relationship between data, and hence such fields are always set to be nullable.
For relationships generated from foreign key constraints (corresponding to `foreign_key_constraint_on`), we distinguish between two cases:
1. The "forward" object relationship from a referencing table (i.e. which has the foreign key constraint) to a referenced table. This should be set to be non-nullable when all referencing columns are non-nullable. But in fact, it used to set it to be non-nullable if *any* referencing column is non-nullable, which is only correct in Postgres when `MATCH FULL` is set (a flag we don't consider). This fixes that by changing a boolean conjunction to a disjunction.
2. The "reverse" object relationship from a referenced table to a referencing table which has the foreign key constraint. This should always be set to be nullable. But in fact, it used to always be set to non-nullable, as was reported in hasura/graphql-engine#7201. This fixes that.
Moreover, we have moved the computation of the nullability from `Hasura.RQL.DDL.Relationship` to `Hasura.GraphQL.Schema.Select`: this nullability used to be passed through the `riIsNullable` field of `RelInfo`, but for array relationships this information is not actually used, and moreover the remaining fields of `RelInfo` are already enough to deduce the nullability.
This also adds regression tests for both (1) and (2) above.
https://github.com/hasura/graphql-engine-mono/pull/2159
GitOrigin-RevId: 617f12765614f49746d18d3368f41dfae2f3e6ca
>
### Description
>
This PR supersedes https://github.com/hasura/graphql-engine-mono/pull/1484. Apply `limit` to the table selection before joining relationship rows to improve query performance.
### Changelog
- [x] `CHANGELOG.md` is updated with user-facing content relevant to this PR. If no changelog is required, then add the `no-changelog-required` label.
### Affected components
- [x] Server
### Related Issues
->
Fix https://github.com/hasura/graphql-engine/issues/5745
### Solution and Design
>
Prior to this change, we apply `LIMIT` and `OFFSET` to the outer selection from sub-query which includes joins for relationships. Now, we move `LIMIT` and `OFFSET` (if present) to inner selection of base table. But, this isn't done always! If there are order by relationships' columns we apply at the outer selection. To know more, please refer to [source code note](https://github.com/hasura/graphql-engine-mono/pull/2078/files#diff-46d868ee45d3eaac667cebb34731f573c77d5c9c8097bb9ccf1115fc07f65bfdR652).
```graphql
query {
article(limit: 2){
id
title
content
author{
name
}
}
}
```
Before:
```sql
SELECT
coalesce(json_agg("root"), '[]') AS "root"
FROM
(
SELECT
row_to_json(
(
SELECT
"_4_e"
FROM
(
SELECT
"_0_root.base"."id" AS "id",
"_0_root.base"."title" AS "title",
"_0_root.base"."content" AS "content",
"_3_root.or.author"."author" AS "author"
) AS "_4_e"
)
) AS "root"
FROM
(
SELECT
*
FROM
"public"."article"
WHERE
('true')
) AS "_0_root.base"
LEFT OUTER JOIN LATERAL (
SELECT
row_to_json(
(
SELECT
"_2_e"
FROM
(
SELECT
"_1_root.or.author.base"."name" AS "name"
) AS "_2_e"
)
) AS "author"
FROM
(
SELECT
*
FROM
"public"."author"
WHERE
(("_0_root.base"."author_id") = ("id"))
) AS "_1_root.or.author.base"
) AS "_3_root.or.author" ON ('true')
LIMIT
2
) AS "_5_root"
```
cost
```
Aggregate (cost=0.73..0.74 rows=1 width=32)
-> Limit (cost=0.15..0.71 rows=2 width=32)
-> Nested Loop Left Join (cost=0.15..223.96 rows=810 width=32)
-> Seq Scan on article (cost=0.00..18.10 rows=810 width=72)
-> Index Scan using author_pkey on author (cost=0.15..0.24 rows=1 width=36)
Index Cond: (article.author_id = id)
SubPlan 1
-> Result (cost=0.00..0.01 rows=1 width=32)
SubPlan 2
-> Result (cost=0.00..0.01 rows=1 width=32)
```
After:
```sql
SELECT
coalesce(json_agg("root"), '[]') AS "root"
FROM
(
SELECT
row_to_json(
(
SELECT
"_4_e"
FROM
(
SELECT
"_0_root.base"."id" AS "id",
"_0_root.base"."title" AS "title",
"_0_root.base"."content" AS "content",
"_3_root.or.author"."author" AS "author"
) AS "_4_e"
)
) AS "root"
FROM
(
SELECT
*
FROM
"public"."article"
WHERE
('true')
LIMIT
2
) AS "_0_root.base"
LEFT OUTER JOIN LATERAL (
SELECT
row_to_json(
(
SELECT
"_2_e"
FROM
(
SELECT
"_1_root.or.author.base"."name" AS "name"
) AS "_2_e"
)
) AS "author"
FROM
(
SELECT
*
FROM
"public"."author"
WHERE
(("_0_root.base"."author_id") = ("id"))
) AS "_1_root.or.author.base"
) AS "_3_root.or.author" ON ('true')
) AS "_5_root"
```
cost:
```
Aggregate (cost=16.47..16.48 rows=1 width=32)
-> Nested Loop Left Join (cost=0.15..16.44 rows=2 width=100)
-> Limit (cost=0.00..0.04 rows=2 width=72)
-> Seq Scan on article (cost=0.00..18.10 rows=810 width=72)
-> Index Scan using author_pkey on author (cost=0.15..8.18 rows=1 width=36)
Index Cond: (article.author_id = id)
SubPlan 1
-> Result (cost=0.00..0.01 rows=1 width=32)
SubPlan 2
-> Result (cost=0.00..0.01 rows=1 width=32)
```
https://github.com/hasura/graphql-engine-mono/pull/2078
Co-authored-by: Evie Ciobanu <1017953+eviefp@users.noreply.github.com>
GitOrigin-RevId: 47eaccdbfb3499efd2c9f733f3312ad31c77916f
The `LazyTxT` type was introduced to avoid connecting to Postgres when a given GraphQL request did not require this. However, through the new way query execution plans are represented, this has now _mostly_ been taken care of in a different way, and so no `LazyTxT` action is generated at all for GraphQL requests that do not fetch data from Postgres.
This removes the laziness of `LazyTxT` by simply making it a newtype wrapper around `Q.TxET`. This simplifies a lot of code in `Hasura.Backends.Postgres.Connection`.
https://github.com/hasura/graphql-engine-mono/pull/1959
GitOrigin-RevId: 58b4d5a05d67bc602b59e02ac338f1c3e63859c7
Query plan caching was introduced by - I believe - hasura/graphql-engine#1934 in order to reduce the query response latency. During the development of PDV in hasura/graphql-engine#4111, it was found out that the new architecture (for which query plan caching wasn't implemented) performed comparably to the pre-PDV architecture with caching. Hence, it was decided to leave query plan caching until some day in the future when it was deemed necessary.
Well, we're in the future now, and there still isn't a convincing argument for query plan caching. So the time has come to remove some references to query plan caching from the codebase. For the most part, any code being removed would probably not be very well suited to the post-PDV architecture of query execution, so arguably not much is lost.
Apart from simplifying the code, this PR will contribute towards making the GraphQL schema generation more modular, testable, and easier to profile. I'd like to eventually work towards a situation in which it's easy to generate a GraphQL schema parser *in isolation*, without being connected to a database, and then parse a GraphQL query *in isolation*, without even listening any HTTP port. It is important that both of these operations can be examined in detail, and in isolation, since they are two major performance bottlenecks, as well as phases where many important upcoming features hook into.
Implementation
The following have been removed:
- The entirety of `server/src-lib/Hasura/GraphQL/Execute/Plan.hs`
- The core phases of query parsing and execution no longer have any references to query plan caching. Note that this is not to be confused with query *response* caching, which is not affected by this PR. This includes removal of the types:
- - `Opaque`, which is replaced by a tuple. Note that the old implementation was broken and did not adequately hide the constructors.
- - `QueryReusability` (and the `markNotReusable` method). Notably, the implementation of the `ParseT` monad now consists of two, rather than three, monad transformers.
- Cache-related tests (in `server/src-test/Hasura/CacheBoundedSpec.hs`) have been removed .
- References to query plan caching in the documentation.
- The `planCacheOptions` in the `TenantConfig` type class was removed. However, during parsing, unrecognized fields in the YAML config get ignored, so this does not cause a breaking change. (Confirmed manually, as well as in consultation with @sordina.)
- The metrics no longer send cache hit/miss messages.
There are a few places in which one can still find references to query plan caching:
- We still accept the `--query-plan-cache-size` command-line option for backwards compatibility. The `HASURA_QUERY_PLAN_CACHE_SIZE` environment variable is not read.
https://github.com/hasura/graphql-engine-mono/pull/1815
GitOrigin-RevId: 17d92b254ec093c62a7dfeec478658ede0813eb7
GJ IR changes cherry-picked from the original GJ branch. There is a separate (can be merged independently) PR for metadata changes (#1727) and there will be a different PR upcoming PR for execution changes.
https://github.com/hasura/graphql-engine-mono/pull/1810
Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com>
GitOrigin-RevId: c31956af29dc9c9b75d002aba7d93c230697c5f4
## Description
This PR fixes an oversight in the implementation of the resolvers of different backends. To implement resolution from environment variables, both MSSQL and BigQuery were directly fetching the process' environment variables, instead of using the careful curated set we thread from main. It was working just fine on OSS, but is failing on Cloud.
This PR fixes this by adding an additional argument to `resolveSourceConfig`, to ensure that backends always use the correct set of variables.
https://github.com/hasura/graphql-engine-mono/pull/1891
GitOrigin-RevId: 58644cab7d041a8bf4235e2acfe9cf71533a92a1
### Description
This PR is the first of several PRs meant to introduce Generalized Joins. In this first PR, we add non-breaking changes to the Metadata types for DB-to-DB remote joins. Note that we are currently rejecting the new remote join format in order to keep folks from breaking their metadata (in case of a downgrade). These issues will be tackled (and JSON changes reverted) in subsequent PRs.
This PR also changes the way we construct the schema cache, and breaks the way we process sources in two steps: we first resolve each source and construct a cache of their tables' raw info, then in a second step we build the source output. This is so that we have access to the target source's tables when building db-to-db relationships.
### Notes
- this PR contains a few minor cleanups of the schema
- it also fixes a bug in how we do renames in remote schema relationships
- it introduces cross-source schema dependencies
https://github.com/hasura/graphql-engine-mono/pull/1727
Co-authored-by: Evie Ciobanu <1017953+eviefp@users.noreply.github.com>
GitOrigin-RevId: f625473077bc5fff5d941b70e9a116192bc1eb22
### Description
In our haste to generalize everything for MSSQL, we put every single "suspicious" type in Backend, including ones that weren't required. `Alias` is one of those: it's only used in a type alias, and is actually just an implementation detail of the translation layer. This PR removes it.
https://github.com/hasura/graphql-engine-mono/pull/1759
GitOrigin-RevId: fb348934ec65a51aae7f95d93c83c3bb704587b5
### Description
This PR removes all `fmapX` and `traverseX` functions from RQL.IR, favouring instead `Functor` and `Traversable` instances throughout the code. This was a relatively straightforward change, except for two small pain points: `AnnSelectG` and `AnnInsert`. Both were parametric over two types `a` and `v`, making it impossible to make them traversable functors... But it turns out that in every single use case, `a ~ f v`. By changing those types to take such an `f :: Type -> Type` as an argument instead of `a :: Type` makes it possible to make them functors.
The only small difference is for `AnnIns`, I had to introduce one `Identity` transformation for one of the `f` parameters. This is relatively straightforward.
### Notes
This PR fixes the most verbose BigQuery hint (`let` instead of `<- pure`).
https://github.com/hasura/graphql-engine-mono/pull/1668
GitOrigin-RevId: e632263a8c559aa04aeae10dcaec915b4a81ad1a
* fix resetting the catalog version to 43 on migration from 1.0 to 2.0
* ci: remove applying patch in test_oss_server_upgrade job
* make the 43 to 46th migrations idempotent
* Set missing HASURA_GRAPHQL_EVENTS_HTTP_POOL_SIZE=8 in upgrade_test
It's not clear why this wasn't caught in CI.
* ci: disable one component of event backpressure test
Co-authored-by: Vishnu Bharathi P <vishnubharathi04@gmail.com>
Co-authored-by: Karthikeyan Chinnakonda <karthikeyan@hasura.io>
Co-authored-by: Brandon Simmons <brandon@hasura.io>
GitOrigin-RevId: c74c6425266a99165c6beecc3e4f7c34e6884d4d
### Description
This PR adds the required IR for DB to DB joins, based on @paf31 and @0x777 's `feature/db-to-db` branch.
To do so, it also refactors the IR to introduce a new type parameter, `r`, which is used to recursively constructs the `v` parameter of remote QueryDBs. When collecting remote joins, we replace `r` with `Const Void`, indicating at the type level that there cannot be any leftover remote join.
Furthermore, this PR refactors IR.Select for readability, moves some code from IR.Root to IR.Select to avoid having to deal with circular dependencies, and makes it compile by adding `error` in all new cases in the execution pipeline.
The diff doesn't make it clear, but most of Select.hs is actually unchanged. Declarations have just been reordered by topic, in the following order:
- type declarations
- instance declarations
- type aliases
- constructor functions
- traverse functions
https://github.com/hasura/graphql-engine-mono/pull/1580
Co-authored-by: Phil Freeman <630306+paf31@users.noreply.github.com>
GitOrigin-RevId: bbdcb4119cec8bb3fc32f1294f91b8dea0728721
Remote relationships are now supported on SQL Server and BigQuery. The major change though is the re-architecture of remote join execution logic. Prior to this PR, each backend is responsible for processing the remote relationships that are part of their AST.
This is not ideal as there is nothing specific about a remote join's execution that ties it to a backend. The only backend specific part is whether or not the specification of the remote relationship is valid (i.e, we'll need to validate whether the scalars are compatible).
The approach now changes to this:
1. Before delegating the AST to the backend, we traverse the AST, collect all the remote joins while modifying the AST to add necessary join fields where needed.
1. Once the remote joins are collected from the AST, the database call is made to fetch the response. The necessary data for the remote join(s) is collected from the database's response and one or more remote schema calls are constructed as necessary.
1. The remote schema calls are then executed and the data from the database and from the remote schemas is joined to produce the final response.
### Known issues
1. Ideally the traversal of the IR to collect remote joins should return an AST which does not include remote join fields. This operation can be type safe but isn't taken up as part of the PR.
1. There is a lot of code duplication between `Transport/HTTP.hs` and `Transport/Websocket.hs` which needs to be fixed ASAP. This too hasn't been taken up by this PR.
1. The type which represents the execution plan is only modified to handle our current remote joins and as such it will have to be changed to accommodate general remote joins.
1. Use of lenses would have reduced the boilerplate code to collect remote joins from the base AST.
1. The current remote join logic assumes that the join columns of a remote relationship appear with their names in the database response. This however is incorrect as they could be aliased. This can be taken up by anyone, I've left a comment in the code.
### Notes to the reviewers
I think it is best reviewed commit by commit.
1. The first one is very straight forward.
1. The second one refactors the remote join execution logic but other than moving things around, it doesn't change the user facing functionality. This moves Postgres specific parts to `Backends/Postgres` module from `Execute`. Some IR related code to `Hasura.RQL.IR` module. Simplifies various type class function signatures as a backend doesn't have to handle remote joins anymore
1. The third one fixes partial case matches that for some weird reason weren't shown as warnings before this refactor
1. The fourth one generalizes the validation logic of remote relationships and implements `scalarTypeGraphQLName` function on SQL Server and BigQuery which is used by the validation logic. This enables remote relationships on BigQuery and SQL Server.
https://github.com/hasura/graphql-engine-mono/pull/1497
GitOrigin-RevId: 77dd8eed326602b16e9a8496f52f46d22b795598
This reverts the remote schema type customisation and namespacing feature temporarily as we test for certain conditions.
GitOrigin-RevId: f8ee97233da4597f703970c3998664c03582d8e7
event catalog:
- `hdb_catalog` is no longer automatically created
- catalog is initialised when the first event trigger is created
- catalog initialisation is done during the schema cache build, using `ArrowCache` so it is only run in response to a change to the set of event triggers
event queue:
- `processEventQueue` thread is prevented from starting when `HASURA_GRAPHQL_EVENTS_FETCH_INTERVAL=0`
- `processEventQueue` thread only processes sources for which at least one event trigger exists in some table in the source
Co-authored-by: Anon Ray <616387+ecthiender@users.noreply.github.com>
GitOrigin-RevId: 73f256465d62490cd2b59dcd074718679993d4fe
