Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- | Tools to analyze the structure of a GraphQL request.
|
2022-01-21 08:39:08 +03:00
|
|
|
module Hasura.GraphQL.Analyse
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
( -- * Query structure
|
|
|
|
Structure (..),
|
|
|
|
FieldInfo (..),
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
InputFieldInfo (..),
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
VariableInfo (..),
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
ScalarInfo (..),
|
|
|
|
EnumInfo (..),
|
|
|
|
ObjectInfo (..),
|
|
|
|
InputObjectInfo (..),
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
-- * Analysis
|
|
|
|
diagnoseGraphQLQuery,
|
|
|
|
analyzeGraphQLQuery,
|
2022-01-21 08:39:08 +03:00
|
|
|
)
|
|
|
|
where
|
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
import Control.Monad.Circular
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
import Control.Monad.Writer (Writer, runWriter)
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
import Data.HashMap.Strict.Extended qualified as Map
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
import Data.Sequence ((|>))
|
|
|
|
import Data.Text qualified as T
|
2022-06-23 12:14:24 +03:00
|
|
|
import Hasura.GraphQL.Parser.Name qualified as GName
|
|
|
|
import Hasura.Name qualified as Name
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
import Hasura.Prelude
|
2022-01-21 08:39:08 +03:00
|
|
|
import Language.GraphQL.Draft.Syntax qualified as G
|
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
--------------------------------------------------------------------------------
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- GraphQL query structure
|
2022-01-21 08:39:08 +03:00
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
-- | Overall structure of a given query. We extract the tree of fields in the
|
|
|
|
-- output, and the graph of input variables.
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
data Structure = Structure
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
{ _stSelection :: HashMap G.Name FieldInfo,
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
_stVariables :: HashMap G.Name VariableInfo
|
2022-01-21 08:39:08 +03:00
|
|
|
}
|
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
-- | Information about the type of an output field; whether the base type is an
|
|
|
|
-- object or a scalar, we store the correspoding 'GType' to keep track of the
|
|
|
|
-- modifiers applied to it (list or non-nullability).
|
|
|
|
data FieldInfo
|
|
|
|
= FieldObjectInfo G.GType ObjectInfo
|
|
|
|
| FieldScalarInfo G.GType ScalarInfo
|
|
|
|
| FieldEnumInfo G.GType EnumInfo
|
2022-01-21 08:39:08 +03:00
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
data ScalarInfo = ScalarInfo
|
|
|
|
{ _siTypeDefinition :: G.ScalarTypeDefinition
|
|
|
|
}
|
2022-01-21 08:39:08 +03:00
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
data EnumInfo = EnumInfo
|
|
|
|
{ _eiTypeDefinition :: G.EnumTypeDefinition
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
}
|
2022-01-21 08:39:08 +03:00
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
data ObjectInfo = ObjectInfo
|
|
|
|
{ _oiTypeDefinition :: G.ObjectTypeDefinition G.InputValueDefinition,
|
|
|
|
_oiSelection :: HashMap G.Name FieldInfo
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
}
|
2022-01-21 08:39:08 +03:00
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
-- | Information about a single variable of the query.
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
data VariableInfo = VariableInfo
|
|
|
|
{ _viType :: G.GType,
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
_viTypeInfo :: InputFieldInfo,
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
_viDefaultValue :: Maybe (G.Value Void)
|
|
|
|
}
|
2022-01-21 08:39:08 +03:00
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
-- | Information about the type of an input field; whether the base type is an
|
|
|
|
-- object or a scalar, we store the correspoding 'GType' to keep track of the
|
|
|
|
-- modifiers applied to it (list or non-nullability).
|
|
|
|
data InputFieldInfo
|
|
|
|
= InputFieldScalarInfo ScalarInfo
|
|
|
|
| InputFieldEnumInfo EnumInfo
|
|
|
|
| InputFieldObjectInfo InputObjectInfo
|
|
|
|
|
|
|
|
data InputObjectInfo = InputObjectInfo
|
|
|
|
{ _ioiTypeDefinition :: G.InputObjectTypeDefinition G.InputValueDefinition,
|
|
|
|
-- | lazy for knot-tying, as we build a graph
|
|
|
|
_ioiFields :: ~(HashMap G.Name (G.GType, InputFieldInfo))
|
|
|
|
}
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- Analysis
|
2022-03-08 12:48:21 +03:00
|
|
|
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- | Given the schema's definition, and a query, validate that the query is
|
|
|
|
-- consistent. We do this by running the analysis, but discarding the result: we
|
|
|
|
-- do not care about the structure, only about the validity of the query.
|
|
|
|
--
|
|
|
|
-- Returns 'Nothing' if the query is valid, or a list of messages otherwise.
|
|
|
|
diagnoseGraphQLQuery ::
|
Decouple `Analyse` and `OpenAPI` from remote schema introspection and internal execution details.
### Motivation
#2338 introduced a way to validate REST queries against the metadata after a change, to properly report any inconsistency that would emerge from a change in the underlying structure of our schema. However, the way this was done was quite complex and error-prone. Namely: we would use the generated schema parsers to statically execute an introspection query, similar to the one we use for remote schemas, then parse the resulting bytestring as it were coming from a remote schema.
This led to several issues: the code was using remote schema primitives, and was associated with remote schema code, despite being unrelated, which led to absurd situations like creating fake `Variable`s whose type was also their name. A lot of the code had to deal with the fact that we might fail to re-parse our own schema. Additionally, some of it was dead code, that for some reason GHC did not warn about? But more fundamentally, this architecture decision creates a dependency between unrelated pieces of the engine: modifying the internal processing of root fields or the introspection of remote schemas now risks impacting the unrelated `OpenAPI` feature.
### Description
This PR decouples that process from the remote schema introspection logic and from the execution engine by making `Analyse` and `OpenAPI` work on the generic `G.SchemaIntrospection` instead. To accomplish this, it:
- adds `GraphQL.Parser.Schema.Convert`, to convert from our "live" schema back to a flat `SchemaIntrospection`
- persists in the schema cache the `admin` introspection generated when building the schema, and uses it both for validation and for generating the `OpenAPI`.
### Known issues and limitations
This adds a bit of memory pressure to the engine, as we persist the entire schema in the schema cache. This might be acceptable in the short-term, but we have several potential ideas going forward should this be a problem:
- cache the result of `Analyze`: when it becomes possible to build the `OpenAPI` purely with the result of `Analyze` without any additional schema information, then we could cache that instead, reducing the footprint
- caching the `OpenAPI`: if it doesn't need to change every time the endpoint is queried, then it should be possible to cache the entire `OpenAPI` object instead of the schema
- cache a copy of the `FieldParsers` used to generate the schema: as those are persisted through the GraphQL `Context`, and are the only input required to generate the `Schema`, making them accessible in the schema cache would allow us to have the exact same feature with no additional memory cost, at the price of a slightly slower and more complicated process (need to rebuild the `Schema` every time we query the OpenAPI endpoint)
- cache nothing at all, and rebuild the admin schema from scratch every time.
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3962
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: a8b9808170b231fdf6787983b4a9ed286cde27e0
2022-03-22 10:36:39 +03:00
|
|
|
G.SchemaIntrospection ->
|
2022-05-04 13:56:54 +03:00
|
|
|
G.TypedOperationDefinition G.NoFragments G.Name ->
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
Maybe [Text]
|
|
|
|
diagnoseGraphQLQuery schema query =
|
|
|
|
let (_structure, errors) = analyzeGraphQLQuery schema query
|
|
|
|
in if null errors
|
|
|
|
then Nothing
|
|
|
|
else Just errors
|
|
|
|
|
|
|
|
-- | Given the schema's definition, and a query, run the analysis.
|
|
|
|
--
|
|
|
|
-- We process all possible fields, and return a partially filled structure if
|
|
|
|
-- necessary. Given the following query:
|
|
|
|
--
|
|
|
|
-- > query {
|
|
|
|
-- > foo {
|
|
|
|
-- > bar
|
|
|
|
-- > }
|
|
|
|
-- > does_not_exist {
|
|
|
|
-- > ghsdflgh
|
|
|
|
-- > }
|
|
|
|
-- > }
|
|
|
|
--
|
|
|
|
-- We would return a structure containing:
|
|
|
|
--
|
|
|
|
-- > foo: {
|
|
|
|
-- > bar: {
|
|
|
|
-- > }
|
|
|
|
-- > }
|
|
|
|
--
|
|
|
|
-- AND an error about "does_not_exist" not existing.
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
--
|
|
|
|
-- In some cases, however, we might not be able to produce a structure at all,
|
|
|
|
-- in which case we return 'Nothing'. This either indicates that something was
|
|
|
|
-- fundamentally wrong with the structure of the query (such as not finding an
|
|
|
|
-- object at the top level), or that a recoverable error was not caught properly
|
|
|
|
-- (see 'withCatchAndRecord').
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
analyzeGraphQLQuery ::
|
Decouple `Analyse` and `OpenAPI` from remote schema introspection and internal execution details.
### Motivation
#2338 introduced a way to validate REST queries against the metadata after a change, to properly report any inconsistency that would emerge from a change in the underlying structure of our schema. However, the way this was done was quite complex and error-prone. Namely: we would use the generated schema parsers to statically execute an introspection query, similar to the one we use for remote schemas, then parse the resulting bytestring as it were coming from a remote schema.
This led to several issues: the code was using remote schema primitives, and was associated with remote schema code, despite being unrelated, which led to absurd situations like creating fake `Variable`s whose type was also their name. A lot of the code had to deal with the fact that we might fail to re-parse our own schema. Additionally, some of it was dead code, that for some reason GHC did not warn about? But more fundamentally, this architecture decision creates a dependency between unrelated pieces of the engine: modifying the internal processing of root fields or the introspection of remote schemas now risks impacting the unrelated `OpenAPI` feature.
### Description
This PR decouples that process from the remote schema introspection logic and from the execution engine by making `Analyse` and `OpenAPI` work on the generic `G.SchemaIntrospection` instead. To accomplish this, it:
- adds `GraphQL.Parser.Schema.Convert`, to convert from our "live" schema back to a flat `SchemaIntrospection`
- persists in the schema cache the `admin` introspection generated when building the schema, and uses it both for validation and for generating the `OpenAPI`.
### Known issues and limitations
This adds a bit of memory pressure to the engine, as we persist the entire schema in the schema cache. This might be acceptable in the short-term, but we have several potential ideas going forward should this be a problem:
- cache the result of `Analyze`: when it becomes possible to build the `OpenAPI` purely with the result of `Analyze` without any additional schema information, then we could cache that instead, reducing the footprint
- caching the `OpenAPI`: if it doesn't need to change every time the endpoint is queried, then it should be possible to cache the entire `OpenAPI` object instead of the schema
- cache a copy of the `FieldParsers` used to generate the schema: as those are persisted through the GraphQL `Context`, and are the only input required to generate the `Schema`, making them accessible in the schema cache would allow us to have the exact same feature with no additional memory cost, at the price of a slightly slower and more complicated process (need to rebuild the `Schema` every time we query the OpenAPI endpoint)
- cache nothing at all, and rebuild the admin schema from scratch every time.
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3962
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: a8b9808170b231fdf6787983b4a9ed286cde27e0
2022-03-22 10:36:39 +03:00
|
|
|
G.SchemaIntrospection ->
|
2022-05-04 13:56:54 +03:00
|
|
|
G.TypedOperationDefinition G.NoFragments G.Name ->
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
(Maybe Structure, [Text])
|
|
|
|
analyzeGraphQLQuery schema G.TypedOperationDefinition {..} = runAnalysis schema do
|
|
|
|
-- analyze the selection
|
|
|
|
selection <- withCatchAndRecord do
|
|
|
|
let rootTypeName = case _todType of
|
|
|
|
G.OperationTypeQuery -> queryRootName
|
|
|
|
G.OperationTypeMutation -> mutationRootName
|
|
|
|
G.OperationTypeSubscription -> subscriptionRootName
|
|
|
|
rootTypeDefinition <-
|
|
|
|
lookupType rootTypeName
|
|
|
|
`onNothingM` throwDiagnosis (TypeNotFound rootTypeName)
|
|
|
|
case rootTypeDefinition of
|
|
|
|
G.TypeDefinitionObject otd ->
|
|
|
|
analyzeObjectSelectionSet otd _todSelectionSet
|
|
|
|
_ ->
|
|
|
|
throwDiagnosis RootTypeNotAnObject
|
|
|
|
-- analyze the variables
|
|
|
|
variables <- analyzeVariables _todVariableDefinitions
|
|
|
|
pure $
|
|
|
|
Structure
|
|
|
|
(fromMaybe mempty selection)
|
|
|
|
variables
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Selection analysis
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
-- | Analyze the fields of an object selection set against its definition, and
|
|
|
|
-- emit the corresponding 'Selection'. We ignore the fields that fail, and we
|
|
|
|
-- continue accumulating the others.
|
|
|
|
analyzeObjectSelectionSet ::
|
|
|
|
G.ObjectTypeDefinition G.InputValueDefinition ->
|
2022-05-04 13:56:54 +03:00
|
|
|
G.SelectionSet G.NoFragments G.Name ->
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
Analysis (HashMap G.Name FieldInfo)
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
analyzeObjectSelectionSet (G.ObjectTypeDefinition {..}) selectionSet = do
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
fields <- traverse analyzeSelection selectionSet
|
|
|
|
foldlM (Map.unionWithM mergeFields) mempty $ catMaybes fields
|
2022-03-08 12:48:21 +03:00
|
|
|
where
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
analyzeSelection :: G.Selection G.NoFragments G.Name -> Analysis (Maybe (HashMap G.Name FieldInfo))
|
|
|
|
analyzeSelection = \case
|
|
|
|
G.SelectionInlineFragment inlineFrag ->
|
|
|
|
mconcat <$> traverse analyzeSelection (G._ifSelectionSet inlineFrag)
|
|
|
|
G.SelectionField field@G.Field {..} ->
|
|
|
|
fmap join $
|
|
|
|
withField _fName $ withCatchAndRecord do
|
|
|
|
-- attempt to find that field in the object's definition
|
|
|
|
G.FieldDefinition {..} <-
|
|
|
|
findDefinition _fName
|
|
|
|
`onNothing` throwDiagnosis (ObjectFieldNotFound _otdName _fName)
|
|
|
|
-- attempt to find its type in the schema
|
|
|
|
let baseType = G.getBaseType _fldType
|
|
|
|
typeDefinition <-
|
|
|
|
lookupType baseType
|
|
|
|
`onNothingM` throwDiagnosis (TypeNotFound baseType)
|
|
|
|
-- attempt to build a corresponding FieldInfo
|
|
|
|
maybeFieldInfo <- analyzeField _fldType typeDefinition field
|
|
|
|
pure $ Map.singleton (fromMaybe _fName _fAlias) <$> maybeFieldInfo
|
|
|
|
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- Additional hidden fields that are allowed despite not being listed in the
|
|
|
|
-- schema.
|
|
|
|
systemFields :: [G.FieldDefinition G.InputValueDefinition]
|
|
|
|
systemFields =
|
|
|
|
if _otdName `elem` [queryRootName, mutationRootName, subscriptionRootName]
|
|
|
|
then [typenameField, schemaField, typeField]
|
|
|
|
else [typenameField]
|
|
|
|
|
|
|
|
-- Search for that field in the schema's definition.
|
|
|
|
findDefinition :: G.Name -> Maybe (G.FieldDefinition G.InputValueDefinition)
|
|
|
|
findDefinition name =
|
|
|
|
find
|
|
|
|
(\fieldDef -> G._fldName fieldDef == name)
|
|
|
|
(_otdFieldsDefinition <> systemFields)
|
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
-- We collect fields in a @Hashmap Name FieldInfo@; in some cases, we might
|
|
|
|
-- end up with two entries with the same name, in the case where a query
|
|
|
|
-- selects the same field twice; when that happens we attempt to gracefully
|
|
|
|
-- merge the info.
|
|
|
|
mergeFields :: G.Name -> FieldInfo -> FieldInfo -> Analysis FieldInfo
|
|
|
|
mergeFields name field1 field2 = case (field1, field2) of
|
|
|
|
-- both are scalars: we check that they're the same
|
|
|
|
(FieldScalarInfo t1 s1, FieldScalarInfo t2 _) -> do
|
|
|
|
when (t1 /= t2) $
|
|
|
|
throwDiagnosis $
|
|
|
|
MismatchedFields name t1 t2
|
|
|
|
pure $ FieldScalarInfo t1 s1
|
|
|
|
-- both are enums: we check that they're the same
|
|
|
|
(FieldEnumInfo t1 e1, FieldEnumInfo t2 _) -> do
|
|
|
|
when (t1 /= t2) $
|
|
|
|
throwDiagnosis $
|
|
|
|
MismatchedFields name t1 t2
|
|
|
|
pure $ FieldEnumInfo t1 e1
|
|
|
|
-- both are objects, we merge their selection sets
|
|
|
|
(FieldObjectInfo t1 o1, FieldObjectInfo t2 o2) -> do
|
|
|
|
when (t1 /= t2) $
|
|
|
|
throwDiagnosis $
|
|
|
|
MismatchedFields name t1 t2
|
|
|
|
mergedSelection <-
|
|
|
|
Map.unionWithM
|
|
|
|
mergeFields
|
|
|
|
(_oiSelection o1)
|
|
|
|
(_oiSelection o2)
|
|
|
|
pure $ FieldObjectInfo t1 o1 {_oiSelection = mergedSelection}
|
|
|
|
-- they do not match
|
|
|
|
_ ->
|
|
|
|
throwDiagnosis $ MismatchedFields name (getFieldType field1) (getFieldType field2)
|
|
|
|
|
|
|
|
-- Extract the GType of a given field
|
|
|
|
getFieldType = \case
|
|
|
|
FieldEnumInfo t _ -> t
|
|
|
|
FieldScalarInfo t _ -> t
|
|
|
|
FieldObjectInfo t _ -> t
|
|
|
|
|
|
|
|
-- | Analyze a given field, and attempt to build a corresponding 'FieldInfo'.
|
|
|
|
analyzeField ::
|
|
|
|
G.GType ->
|
|
|
|
G.TypeDefinition [G.Name] G.InputValueDefinition ->
|
|
|
|
G.Field G.NoFragments G.Name ->
|
|
|
|
Analysis (Maybe FieldInfo)
|
|
|
|
analyzeField gType typeDefinition G.Field {..} = case typeDefinition of
|
|
|
|
G.TypeDefinitionInputObject iotd -> do
|
|
|
|
-- input objects aren't allowed in selection sets
|
|
|
|
throwDiagnosis $ InputObjectInOutput $ G._iotdName iotd
|
|
|
|
G.TypeDefinitionScalar std -> do
|
|
|
|
-- scalars do not admit a selection set
|
|
|
|
unless (null _fSelectionSet) $
|
|
|
|
throwDiagnosis $
|
|
|
|
ScalarSelectionSet $
|
|
|
|
G._stdName std
|
|
|
|
pure $ Just $ FieldScalarInfo gType $ ScalarInfo std
|
|
|
|
G.TypeDefinitionEnum etd -> do
|
|
|
|
-- enums do not admit a selection set
|
|
|
|
unless (null _fSelectionSet) $
|
|
|
|
throwDiagnosis $
|
|
|
|
EnumSelectionSet $
|
|
|
|
G._etdName etd
|
|
|
|
pure $ Just $ FieldEnumInfo gType $ EnumInfo etd
|
|
|
|
G.TypeDefinitionUnion _utd ->
|
|
|
|
-- TODO: implement unions
|
|
|
|
pure Nothing
|
|
|
|
G.TypeDefinitionInterface _itd ->
|
|
|
|
-- TODO: implement interfaces
|
|
|
|
pure Nothing
|
|
|
|
G.TypeDefinitionObject otd -> do
|
|
|
|
-- TODO: check field arguments?
|
|
|
|
when (null _fSelectionSet) $
|
|
|
|
throwDiagnosis $
|
|
|
|
ObjectMissingSelectionSet $
|
|
|
|
G._otdName otd
|
|
|
|
subselection <- analyzeObjectSelectionSet otd _fSelectionSet
|
|
|
|
pure $
|
|
|
|
Just $
|
|
|
|
FieldObjectInfo gType $
|
|
|
|
ObjectInfo
|
|
|
|
{ _oiTypeDefinition = otd,
|
|
|
|
_oiSelection = subselection
|
|
|
|
}
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Variables analysis
|
|
|
|
|
|
|
|
-- | Analyzes the variables in the given query. This builds the graph of input
|
|
|
|
-- types associated with the variable. This process is, like any GraphQL schema
|
|
|
|
-- operation, inherently self-recursive, and we use 'CircularT' (a lesser
|
|
|
|
-- 'SchemaT') to tie the knot.
|
|
|
|
analyzeVariables ::
|
|
|
|
[G.VariableDefinition] ->
|
|
|
|
Analysis (HashMap G.Name VariableInfo)
|
|
|
|
analyzeVariables variables = do
|
|
|
|
result <- runCircularT $ for variables \G.VariableDefinition {..} -> do
|
|
|
|
-- TODO: do we want to differentiate field from variable in the error path?
|
|
|
|
withField _vdName $ withCatchAndRecord do
|
|
|
|
let baseType = G.getBaseType _vdType
|
|
|
|
typeDefinition <-
|
|
|
|
lookupType baseType
|
|
|
|
`onNothingM` throwDiagnosis (TypeNotFound baseType)
|
|
|
|
ifInfo <- analyzeInputField baseType typeDefinition
|
|
|
|
pure $ Map.singleton _vdName $ VariableInfo _vdType ifInfo _vdDefaultValue
|
|
|
|
pure $ fold $ catMaybes result
|
|
|
|
|
|
|
|
-- | Builds an 'InputFieldInfo' for a given typename.
|
|
|
|
--
|
|
|
|
-- This function is "memoized" using 'withCircular' to prevent processing the
|
|
|
|
-- same type more than once in case the input types are self-recursive.
|
|
|
|
analyzeInputField ::
|
|
|
|
G.Name ->
|
|
|
|
G.TypeDefinition [G.Name] G.InputValueDefinition ->
|
|
|
|
CircularT G.Name InputFieldInfo Analysis InputFieldInfo
|
|
|
|
analyzeInputField typeName typeDefinition =
|
|
|
|
withCircular typeName $ case typeDefinition of
|
|
|
|
G.TypeDefinitionScalar std ->
|
|
|
|
pure $ InputFieldScalarInfo (ScalarInfo std)
|
|
|
|
G.TypeDefinitionEnum etd ->
|
|
|
|
pure $ InputFieldEnumInfo (EnumInfo etd)
|
|
|
|
G.TypeDefinitionInputObject iotd -> do
|
|
|
|
fields <- for (G._iotdValueDefinitions iotd) \G.InputValueDefinition {..} -> do
|
|
|
|
withField _ivdName $ withCatchAndRecord do
|
|
|
|
let baseType = G.getBaseType _ivdType
|
|
|
|
typeDef <-
|
|
|
|
lookupType baseType
|
|
|
|
`onNothingM` throwDiagnosis (TypeNotFound baseType)
|
|
|
|
info <- analyzeInputField baseType typeDef
|
|
|
|
pure (_ivdName, (_ivdType, info))
|
|
|
|
pure $ InputFieldObjectInfo (InputObjectInfo iotd $ Map.fromList $ catMaybes fields)
|
|
|
|
G.TypeDefinitionObject _otd -> throwDiagnosis $ ObjectInInput typeName
|
|
|
|
G.TypeDefinitionInterface _itd -> throwDiagnosis $ InterfaceInInput typeName
|
|
|
|
G.TypeDefinitionUnion _utd -> throwDiagnosis $ UnionInInput typeName
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- Internal Analysis monad and helpers
|
|
|
|
|
|
|
|
-- | The monad in which we run our analysis.
|
|
|
|
--
|
|
|
|
-- Has three capabilities:
|
|
|
|
-- - reader carries the current path, and the full schema for lookups
|
|
|
|
-- - writer logs all errors we have caught
|
|
|
|
-- - except allows for short-circuiting errors
|
|
|
|
newtype Analysis a
|
|
|
|
= Analysis
|
|
|
|
( ExceptT
|
|
|
|
AnalysisError
|
|
|
|
( ReaderT
|
|
|
|
(Path, G.SchemaIntrospection)
|
|
|
|
(Writer [AnalysisError])
|
|
|
|
)
|
|
|
|
a
|
|
|
|
)
|
|
|
|
deriving newtype
|
|
|
|
( Functor,
|
|
|
|
Applicative,
|
|
|
|
Monad,
|
|
|
|
MonadReader (Path, G.SchemaIntrospection),
|
|
|
|
MonadWriter [AnalysisError],
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
MonadError AnalysisError,
|
|
|
|
MonadFix
|
2022-01-21 08:39:08 +03:00
|
|
|
)
|
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
runAnalysis :: G.SchemaIntrospection -> Analysis a -> (Maybe a, [Text])
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
runAnalysis schema (Analysis a) =
|
|
|
|
postProcess $
|
|
|
|
runWriter $
|
|
|
|
flip runReaderT (pure "$", schema) $
|
|
|
|
runExceptT a
|
2022-03-08 12:48:21 +03:00
|
|
|
where
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- if there was an uncaught error, add it to the list
|
|
|
|
postProcess = \case
|
|
|
|
(Left err, errors) ->
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
(Nothing, map render $ errors ++ [err])
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
(Right result, errors) ->
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
(Just result, map render errors)
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
-- | Look up a type in the schema.
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
lookupType ::
|
|
|
|
(MonadReader (Path, G.SchemaIntrospection) m) =>
|
|
|
|
G.Name ->
|
|
|
|
m (Maybe (G.TypeDefinition [G.Name] G.InputValueDefinition))
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
lookupType name = do
|
|
|
|
G.SchemaIntrospection types <- asks snd
|
|
|
|
pure $ Map.lookup name types
|
|
|
|
|
|
|
|
-- | Add the current field to the error path.
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
withField ::
|
|
|
|
(MonadReader (Path, G.SchemaIntrospection) m) =>
|
|
|
|
G.Name ->
|
|
|
|
m a ->
|
|
|
|
m a
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
withField name = local $ first (|> G.unName name)
|
|
|
|
|
|
|
|
-- | Throws an 'AnalysisError' by combining the given diagnosis with the current
|
|
|
|
-- path. This interrupts the computation in the given branch, and must be caught
|
|
|
|
-- for the analysis to resume.
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
throwDiagnosis ::
|
|
|
|
( MonadReader (Path, G.SchemaIntrospection) m,
|
|
|
|
MonadError AnalysisError m
|
|
|
|
) =>
|
|
|
|
Diagnosis ->
|
|
|
|
m a
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
throwDiagnosis d = do
|
|
|
|
currentPath <- asks fst
|
|
|
|
throwError $ AnalysisError currentPath d
|
|
|
|
|
|
|
|
-- | Runs the given computation. if it fails, cacthes the error, records it in
|
|
|
|
-- the monad, and return 'Nothing'. This allows for a clean recovery.
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
withCatchAndRecord ::
|
|
|
|
( MonadReader (Path, G.SchemaIntrospection) m,
|
|
|
|
MonadWriter [AnalysisError] m,
|
|
|
|
MonadError AnalysisError m
|
|
|
|
) =>
|
|
|
|
m a ->
|
|
|
|
m (Maybe a)
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
withCatchAndRecord action =
|
|
|
|
fmap Just action `catchError` \e -> do
|
|
|
|
tell [e]
|
|
|
|
pure Nothing
|
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
--------------------------------------------------------------------------------
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- Analysis errors
|
|
|
|
|
|
|
|
data AnalysisError = AnalysisError
|
|
|
|
{ _aePath :: Path,
|
|
|
|
_aeDiagnosis :: Diagnosis
|
|
|
|
}
|
|
|
|
|
|
|
|
type Path = Seq Text
|
|
|
|
|
|
|
|
data Diagnosis
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
= RootTypeNotAnObject
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
| TypeNotFound G.Name
|
|
|
|
| EnumSelectionSet G.Name
|
|
|
|
| ScalarSelectionSet G.Name
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
| InputObjectInOutput G.Name
|
|
|
|
| UnionInInput G.Name
|
|
|
|
| ObjectInInput G.Name
|
|
|
|
| InterfaceInInput G.Name
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
| ObjectFieldNotFound G.Name G.Name
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
| ObjectMissingSelectionSet G.Name
|
|
|
|
| MismatchedFields G.Name G.GType G.GType
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
render :: AnalysisError -> Text
|
|
|
|
render (AnalysisError path diagnosis) =
|
|
|
|
T.intercalate "." (toList path) <> ": " <> case diagnosis of
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
RootTypeNotAnObject ->
|
|
|
|
"the root type is not an object"
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
TypeNotFound name ->
|
|
|
|
"type '" <> G.unName name <> "' not found in the schema"
|
|
|
|
EnumSelectionSet name ->
|
|
|
|
"enum '" <> G.unName name <> "' does not accept a selection set"
|
|
|
|
ScalarSelectionSet name ->
|
|
|
|
"scalar '" <> G.unName name <> "' does not accept a selection set"
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
InputObjectInOutput name ->
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
"input object '" <> G.unName name <> "' cannot be used for output"
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
UnionInInput name ->
|
|
|
|
"union '" <> G.unName name <> "' cannot be used in an input type"
|
|
|
|
ObjectInInput name ->
|
|
|
|
"object '" <> G.unName name <> "' cannot be used in an input type"
|
|
|
|
InterfaceInInput name ->
|
|
|
|
"interface '" <> G.unName name <> "' cannot be used in an input type"
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
ObjectFieldNotFound objName fieldName ->
|
|
|
|
"field '" <> G.unName fieldName <> "' not found in object '" <> G.unName objName <> "'"
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
ObjectMissingSelectionSet objName ->
|
|
|
|
"object of type '" <> G.unName objName <> "' must have a selection set"
|
|
|
|
MismatchedFields name type1 type2 ->
|
|
|
|
"field '" <> G.unName name <> "' seen with two different types: " <> tshow type1 <> " and " <> tshow type2
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
Rewrite OpenAPI
### Description
This PR rewrites OpenAPI to be more idiomatic. Some noteworthy changes:
- we accumulate all required information during the Analyze phase, to avoid having to do a single lookup in the schema cache during the OpenAPI generation phase (we now only need the schema cache as input to run the analysis)
- we no longer build intermediary endpoint information and aggregate it, we directly build the the `PathItem` for each endpoint; additionally, that means we no longer have to assume that different methods have the same metadata
- we no longer have to first declare types, then craft references: we do everything in one step
- we now properly deal with nullability by treating "typeName" and "typeName!" as different
- we add a bunch of additional fields in the generated "schema", such as title
- we do now support enum values in both input and output positions
- checking whether the request body is required is now performed on the fly rather than by introspecting the generated schema
- the methods in the file are sorted by topic
### Controversial point
However, this PR creates some additional complexity, that we might not want to keep. The main complexity is _knot-tying_: to avoid lookups when generating the OpenAPI, it builds an actual graph of input types, which means that we need something similar to (but simpler than) `MonadSchema`, to avoid infinite recursions when analyzing the input types of a query. To do this, this PR introduces `CircularT`, a lesser `SchemaT` that aims at avoiding ever having to reinvent this particular wheel ever again.
### Remaining work
- [x] fix existing tests (they are all failing due to some of the schema changes)
- [ ] add tests to cover the new features:
- [x] tests for `CircularT`
- [ ] tests for enums in output schemas
- [x] extract / document `CircularT` if we wish to keep it
- [x] add more comments to `OpenAPI`
- [x] have a second look at `buildVariableSchema`
- [x] fix all missing diagnostics in `Analyze`
- [x] add a Changelog entry?
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4654
Co-authored-by: David Overton <7734777+dmoverton@users.noreply.github.com>
GitOrigin-RevId: f4a9191f22dfcc1dccefd6a52f5c586b6ad17172
2022-06-30 15:55:56 +03:00
|
|
|
--------------------------------------------------------------------------------
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
-- GraphQL internals
|
|
|
|
|
|
|
|
-- Special type names
|
|
|
|
|
|
|
|
queryRootName :: G.Name
|
2022-06-23 12:14:24 +03:00
|
|
|
queryRootName = Name._query_root
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
mutationRootName :: G.Name
|
2022-06-23 12:14:24 +03:00
|
|
|
mutationRootName = Name._mutation_root
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
subscriptionRootName :: G.Name
|
2022-06-23 12:14:24 +03:00
|
|
|
subscriptionRootName = Name._subscription_root
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
-- Reserved fields
|
|
|
|
|
|
|
|
typenameField :: G.FieldDefinition G.InputValueDefinition
|
2022-06-23 12:14:24 +03:00
|
|
|
typenameField = mkReservedField GName.___typename GName._String
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
schemaField :: G.FieldDefinition G.InputValueDefinition
|
2022-06-23 12:14:24 +03:00
|
|
|
schemaField = mkReservedField GName.___schema GName.___Schema
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
typeField :: G.FieldDefinition G.InputValueDefinition
|
2022-06-23 12:14:24 +03:00
|
|
|
typeField = mkReservedField GName.___type GName.___Type
|
Rewrite `GraphQL.Analysis`
### Motivation
While we strive to write clear code, we have historically struggled at Hasura with having very different styles and standards across the codebase. There's been efforts to standardize our coding style, we have an official styleguide that isn't maintained as closely as it should... We still have some work in front of us.
However, in the last ~year or so, there's been a huge push towards incrementally improving the situation. As part of this we've been blocking PRs that don't add enough comments, or don't improve the files that they touch.
While looking at `Hasura.GraphQL.Analyse`, it became apparent that this file did not meet the engineering standards that I would expect to see addressed during a code review. Some ways in which I think it falls short:
- lack of documentation
- no clear distinction between public / internal components
- "unidiomatic" Haskell code (such as using `Either Result Error`)
While there's no problem with a file looking like this during development, those issues should have been caught at review time. The fact that they weren't indicates a problem in our process that we will need to address: code quality and maintainability is paramount, and we all need to do our part.
### Description
This PR rewrites all of `Hasura.GraphQL.Analyze`, and adapts `Hasura.Server.OpenAPI` accordingly where needed. I've attempted to clarify names and add documentation based on my understanding of the code, and to clean what was unused (such as field variables). I don't think this PR is good enough as is, and I welcome criticism where I got my comments wrong / am happy to help y'all add more.
This PR makes one small change in the way error messages are reported (and adjusts the corresponding test accordingly); each error message is now prefixed with the path within the selection set:
```
⚠️ $.test.foo.bar.baz.mizpelled: field 'mizpelled' not found in object 'Baz'
```
### Note
This PR is currently **on top of #3962**. You can preview the changes in isolation by [diffing the branches](https://github.com/hasura/graphql-engine-mono/compare/nicuveo/clean-rest-endpoint-inconsistency-check..nicuveo/rewrite-analysis).
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3963
Co-authored-by: paritosh-08 <85472423+paritosh-08@users.noreply.github.com>
GitOrigin-RevId: 5ec38e0e753f0c12096a350db0737658495e2f15
2022-04-04 08:53:59 +03:00
|
|
|
|
|
|
|
mkReservedField :: G.Name -> G.Name -> G.FieldDefinition G.InputValueDefinition
|
|
|
|
mkReservedField fieldName typeName =
|
|
|
|
G.FieldDefinition Nothing fieldName [] (G.TypeNamed (G.Nullability False) typeName) []
|