graphql-engine/server/src-lib/Hasura/RQL/DDL/Schema/Cache/Fields.hs

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

371 lines
15 KiB
Haskell
Raw Normal View History

{-# LANGUAGE Arrows #-}
module Hasura.RQL.DDL.Schema.Cache.Fields (addNonColumnFields) where
import Control.Arrow.Extended
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
import Control.Arrow.Interpret
import Control.Lens ((^.), _3, _4)
import Data.Aeson
import Data.Align (align)
import Data.HashMap.Strict.Extended qualified as M
import Data.HashSet qualified as HS
import Data.Sequence qualified as Seq
import Data.Text.Extended
import Data.These (These (..))
import Hasura.Base.Error
import Hasura.Incremental qualified as Inc
import Hasura.Prelude
import Hasura.RQL.DDL.ComputedField
import Hasura.RQL.DDL.Relationship
server: support remote relationships on SQL Server and BigQuery (#1497) Remote relationships are now supported on SQL Server and BigQuery. The major change though is the re-architecture of remote join execution logic. Prior to this PR, each backend is responsible for processing the remote relationships that are part of their AST. This is not ideal as there is nothing specific about a remote join's execution that ties it to a backend. The only backend specific part is whether or not the specification of the remote relationship is valid (i.e, we'll need to validate whether the scalars are compatible). The approach now changes to this: 1. Before delegating the AST to the backend, we traverse the AST, collect all the remote joins while modifying the AST to add necessary join fields where needed. 1. Once the remote joins are collected from the AST, the database call is made to fetch the response. The necessary data for the remote join(s) is collected from the database's response and one or more remote schema calls are constructed as necessary. 1. The remote schema calls are then executed and the data from the database and from the remote schemas is joined to produce the final response. ### Known issues 1. Ideally the traversal of the IR to collect remote joins should return an AST which does not include remote join fields. This operation can be type safe but isn't taken up as part of the PR. 1. There is a lot of code duplication between `Transport/HTTP.hs` and `Transport/Websocket.hs` which needs to be fixed ASAP. This too hasn't been taken up by this PR. 1. The type which represents the execution plan is only modified to handle our current remote joins and as such it will have to be changed to accommodate general remote joins. 1. Use of lenses would have reduced the boilerplate code to collect remote joins from the base AST. 1. The current remote join logic assumes that the join columns of a remote relationship appear with their names in the database response. This however is incorrect as they could be aliased. This can be taken up by anyone, I've left a comment in the code. ### Notes to the reviewers I think it is best reviewed commit by commit. 1. The first one is very straight forward. 1. The second one refactors the remote join execution logic but other than moving things around, it doesn't change the user facing functionality. This moves Postgres specific parts to `Backends/Postgres` module from `Execute`. Some IR related code to `Hasura.RQL.IR` module. Simplifies various type class function signatures as a backend doesn't have to handle remote joins anymore 1. The third one fixes partial case matches that for some weird reason weren't shown as warnings before this refactor 1. The fourth one generalizes the validation logic of remote relationships and implements `scalarTypeGraphQLName` function on SQL Server and BigQuery which is used by the validation logic. This enables remote relationships on BigQuery and SQL Server. https://github.com/hasura/graphql-engine-mono/pull/1497 GitOrigin-RevId: 77dd8eed326602b16e9a8496f52f46d22b795598
2021-06-11 06:26:50 +03:00
import Hasura.RQL.DDL.RemoteRelationship
import Hasura.RQL.DDL.Schema.Cache.Common
import Hasura.RQL.DDL.Schema.Function
import Hasura.RQL.Types.Backend
import Hasura.RQL.Types.Column
import Hasura.RQL.Types.Common
import Hasura.RQL.Types.ComputedField
import Hasura.RQL.Types.Function
import Hasura.RQL.Types.Metadata
import Hasura.RQL.Types.Metadata.Backend
import Hasura.RQL.Types.Metadata.Object
import Hasura.RQL.Types.Relationships.Local
import Hasura.RQL.Types.Relationships.Remote
import Hasura.RQL.Types.SchemaCache
import Hasura.RQL.Types.SchemaCache.Build
import Hasura.RQL.Types.SchemaCacheTypes
import Hasura.RQL.Types.Table
import Hasura.SQL.AnyBackend qualified as AB
import Language.GraphQL.Draft.Syntax qualified as G
addNonColumnFields ::
forall b arr m.
( ArrowChoice arr,
Inc.ArrowDistribute arr,
ArrowWriter (Seq CollectedInfo) arr,
ArrowKleisli m arr,
MonadError QErr m,
BackendMetadata b
) =>
( HashMap SourceName (AB.AnyBackend PartiallyResolvedSource),
SourceName,
HashMap (TableName b) (TableCoreInfoG b (ColumnInfo b) (ColumnInfo b)),
FieldInfoMap (ColumnInfo b),
scaffolding for remote-schemas module The main aim of the PR is: 1. To set up a module structure for 'remote-schemas' package. 2. Move parts by the remote schema codebase into the new module structure to validate it. ## Notes to the reviewer Why a PR with large-ish diff? 1. We've been making progress on the MM project but we don't yet know long it is going to take us to get to the first milestone. To understand this better, we need to figure out the unknowns as soon as possible. Hence I've taken a stab at the first two items in the [end-state](https://gist.github.com/0x777/ca2bdc4284d21c3eec153b51dea255c9) document to figure out the unknowns. Unsurprisingly, there are a bunch of issues that we haven't discussed earlier. These are documented in the 'open questions' section. 1. The diff is large but that is only code moved around and I've added a section that documents how things are moved. In addition, there are fair number of PR comments to help with the review process. ## Changes in the PR ### Module structure Sets up the module structure as follows: ``` Hasura/ RemoteSchema/ Metadata/ Types.hs SchemaCache/ Types.hs Permission.hs RemoteRelationship.hs Build.hs MetadataAPI/ Types.hs Execute.hs ``` ### 1. Types representing metadata are moved Types that capture metadata information (currently scattered across several RQL modules) are moved into `Hasura.RemoteSchema.Metadata.Types`. - This new module only depends on very 'core' modules such as `Hasura.Session` for the notion of roles and `Hasura.Incremental` for `Cacheable` typeclass. - The requirement on database modules is avoided by generalizing the remote schemas metadata to accept an arbitrary 'r' for a remote relationship definition. ### 2. SchemaCache related types and build logic have been moved Types that represent remote schemas information in SchemaCache are moved into `Hasura.RemoteSchema.SchemaCache.Types`. Similar to `H.RS.Metadata.Types`, this module depends on 'core' modules except for `Hasura.GraphQL.Parser.Variable`. It has something to do with remote relationships but I haven't spent time looking into it. The validation of 'remote relationships to remote schema' is also something that needs to be looked at. Rips out the logic that builds remote schema's SchemaCache information from the monolithic `buildSchemaCacheRule` and moves it into `Hasura.RemoteSchema.SchemaCache.Build`. Further, the `.SchemaCache.Permission` and `.SchemaCache.RemoteRelationship` have been created from existing modules that capture schema cache building logic for those two components. This was a fair amount of work. On main, currently remote schema's SchemaCache information is built in two phases - in the first phase, 'permissions' and 'remote relationships' are ignored and in the second phase they are filled in. While remote relationships can only be resolved after partially resolving sources and other remote schemas, the same isn't true for permissions. Further, most of the work that is done to resolve remote relationships can be moved to the first phase so that the second phase can be a very simple traversal. This is the approach that was taken - resolve permissions and as much as remote relationships information in the first phase. ### 3. Metadata APIs related types and build logic have been moved The types that represent remote schema related metadata APIs and the execution logic have been moved to `Hasura.RemoteSchema.MetadataAPI.Types` and `.Execute` modules respectively. ## Open questions: 1. `Hasura.RemoteSchema.Metadata.Types` is so called because I was hoping that all of the metadata related APIs of remote schema can be brought in at `Hasura.RemoteSchema.Metadata.API`. However, as metadata APIs depended on functions from `SchemaCache` module (see [1](https://github.com/hasura/graphql-engine-mono/blob/ceba6d62264603ee5d279814677b29bcc43ecaea/server/src-lib/Hasura/RQL/DDL/RemoteSchema.hs#L55) and [2](https://github.com/hasura/graphql-engine-mono/blob/ceba6d62264603ee5d279814677b29bcc43ecaea/server/src-lib/Hasura/RQL/DDL/RemoteSchema.hs#L91), it made more sense to create a separate top-level module for `MetadataAPI`s. Maybe we can just have `Hasura.RemoteSchema.Metadata` and get rid of the extra nesting or have `Hasura.RemoteSchema.Metadata.{Core,Permission,RemoteRelationship}` if we want to break them down further. 1. `buildRemoteSchemas` in `H.RS.SchemaCache.Build` has the following type: ```haskell buildRemoteSchemas :: ( ArrowChoice arr, Inc.ArrowDistribute arr, ArrowWriter (Seq CollectedInfo) arr, Inc.ArrowCache m arr, MonadIO m, HasHttpManagerM m, Inc.Cacheable remoteRelationshipDefinition, ToJSON remoteRelationshipDefinition, MonadError QErr m ) => Env.Environment -> ( (Inc.Dependency (HashMap RemoteSchemaName Inc.InvalidationKey), OrderedRoles), [RemoteSchemaMetadataG remoteRelationshipDefinition] ) `arr` HashMap RemoteSchemaName (PartiallyResolvedRemoteSchemaCtxG remoteRelationshipDefinition, MetadataObject) ``` Note the dependence on `CollectedInfo` which is defined as ```haskell data CollectedInfo = CIInconsistency InconsistentMetadata | CIDependency MetadataObject -- ^ for error reporting on missing dependencies SchemaObjId SchemaDependency deriving (Eq) ``` this pretty much means that remote schemas is dependent on types from databases, actions, .... How do we fix this? Maybe introduce a typeclass such as `ArrowCollectRemoteSchemaDependencies` which is defined in `Hasura.RemoteSchema` and then implemented in graphql-engine? 1. The dependency on `buildSchemaCacheFor` in `.MetadataAPI.Execute` which has the following signature: ```haskell buildSchemaCacheFor :: (QErrM m, CacheRWM m, MetadataM m) => MetadataObjId -> MetadataModifier -> ``` This can be easily resolved if we restrict what the metadata APIs are allowed to do. Currently, they operate in an unfettered access to modify SchemaCache (the `CacheRWM` constraint): ```haskell runAddRemoteSchema :: ( QErrM m, CacheRWM m, MonadIO m, HasHttpManagerM m, MetadataM m, Tracing.MonadTrace m ) => Env.Environment -> AddRemoteSchemaQuery -> m EncJSON ``` This should instead be changed to restrict remote schema APIs to only modify remote schema metadata (but has access to the remote schemas part of the schema cache), this dependency is completely removed. ```haskell runAddRemoteSchema :: ( QErrM m, MonadIO m, HasHttpManagerM m, MonadReader RemoteSchemasSchemaCache m, MonadState RemoteSchemaMetadata m, Tracing.MonadTrace m ) => Env.Environment -> AddRemoteSchemaQuery -> m RemoteSchemeMetadataObjId ``` The idea is that the core graphql-engine would call these functions and then call `buildSchemaCacheFor`. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6291 GitOrigin-RevId: 51357148c6404afe70219afa71bd1d59bdf4ffc6
2022-10-21 06:13:07 +03:00
PartiallyResolvedRemoteSchemaMap,
DBFunctionsMetadata b,
NonColumnTableInputs b
)
`arr` FieldInfoMap (FieldInfo b)
addNonColumnFields =
proc
( allSources,
source,
server: fix the nullability of object relationships (fix hasura/graphql-engine#7201) When adding object relationships, we set the nullability of the generated GraphQL field based on whether the database backend enforces that the referenced data always exists. For manual relationships (corresponding to `manual_configuration`), the database backend is unaware of any relationship between data, and hence such fields are always set to be nullable. For relationships generated from foreign key constraints (corresponding to `foreign_key_constraint_on`), we distinguish between two cases: 1. The "forward" object relationship from a referencing table (i.e. which has the foreign key constraint) to a referenced table. This should be set to be non-nullable when all referencing columns are non-nullable. But in fact, it used to set it to be non-nullable if *any* referencing column is non-nullable, which is only correct in Postgres when `MATCH FULL` is set (a flag we don't consider). This fixes that by changing a boolean conjunction to a disjunction. 2. The "reverse" object relationship from a referenced table to a referencing table which has the foreign key constraint. This should always be set to be nullable. But in fact, it used to always be set to non-nullable, as was reported in hasura/graphql-engine#7201. This fixes that. Moreover, we have moved the computation of the nullability from `Hasura.RQL.DDL.Relationship` to `Hasura.GraphQL.Schema.Select`: this nullability used to be passed through the `riIsNullable` field of `RelInfo`, but for array relationships this information is not actually used, and moreover the remaining fields of `RelInfo` are already enough to deduce the nullability. This also adds regression tests for both (1) and (2) above. https://github.com/hasura/graphql-engine-mono/pull/2159 GitOrigin-RevId: 617f12765614f49746d18d3368f41dfae2f3e6ca
2021-08-26 18:26:43 +03:00
rawTableInfo,
columns,
remoteSchemaMap,
pgFunctions,
NonColumnTableInputs {..}
)
-> do
objectRelationshipInfos <-
buildInfoMapPreservingMetadata
server: fix the nullability of object relationships (fix hasura/graphql-engine#7201) When adding object relationships, we set the nullability of the generated GraphQL field based on whether the database backend enforces that the referenced data always exists. For manual relationships (corresponding to `manual_configuration`), the database backend is unaware of any relationship between data, and hence such fields are always set to be nullable. For relationships generated from foreign key constraints (corresponding to `foreign_key_constraint_on`), we distinguish between two cases: 1. The "forward" object relationship from a referencing table (i.e. which has the foreign key constraint) to a referenced table. This should be set to be non-nullable when all referencing columns are non-nullable. But in fact, it used to set it to be non-nullable if *any* referencing column is non-nullable, which is only correct in Postgres when `MATCH FULL` is set (a flag we don't consider). This fixes that by changing a boolean conjunction to a disjunction. 2. The "reverse" object relationship from a referenced table to a referencing table which has the foreign key constraint. This should always be set to be nullable. But in fact, it used to always be set to non-nullable, as was reported in hasura/graphql-engine#7201. This fixes that. Moreover, we have moved the computation of the nullability from `Hasura.RQL.DDL.Relationship` to `Hasura.GraphQL.Schema.Select`: this nullability used to be passed through the `riIsNullable` field of `RelInfo`, but for array relationships this information is not actually used, and moreover the remaining fields of `RelInfo` are already enough to deduce the nullability. This also adds regression tests for both (1) and (2) above. https://github.com/hasura/graphql-engine-mono/pull/2159 GitOrigin-RevId: 617f12765614f49746d18d3368f41dfae2f3e6ca
2021-08-26 18:26:43 +03:00
(_rdName . (^. _3))
(\(s, t, c) -> mkRelationshipMetadataObject @b ObjRel (s, t, c))
buildObjectRelationship
-<
(_tciForeignKeys <$> rawTableInfo, map (source,_nctiTable,) _nctiObjectRelationships)
arrayRelationshipInfos <-
buildInfoMapPreservingMetadata
(_rdName . (^. _3))
(mkRelationshipMetadataObject @b ArrRel)
buildArrayRelationship
-<
(_tciForeignKeys <$> rawTableInfo, map (source,_nctiTable,) _nctiArrayRelationships)
let relationshipInfos = objectRelationshipInfos <> arrayRelationshipInfos
computedFieldInfos <-
buildInfoMapPreservingMetadata
(_cfmName . (^. _4))
(\(s, _, t, c) -> mkComputedFieldMetadataObject (s, t, c))
( proc ((a, b), (c, d, e, f)) -> do
o <- interpretWriter -< buildComputedField a b c d e f
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
arrM liftEither -< o
)
-<
( ( HS.fromList $ M.keys rawTableInfo,
HS.fromList $ map ciColumn $ M.elems columns
),
map (source,pgFunctions,_nctiTable,) _nctiComputedFields
)
-- the fields that can be used for defining join conditions to other sources/remote schemas:
-- 1. all columns
-- 2. computed fields which don't expect arguments other than the table row and user session
let lhsJoinFields =
let columnFields = columns <&> \columnInfo -> JoinColumn (ciColumn columnInfo) (ciType columnInfo)
computedFields = M.fromList $
flip mapMaybe (M.toList computedFieldInfos) $
\(cfName, (ComputedFieldInfo {..}, _)) -> do
scalarType <- case computedFieldReturnType @b _cfiReturnType of
ReturnsScalar ty -> pure ty
ReturnsTable {} -> Nothing
ReturnsOthers {} -> Nothing
let ComputedFieldFunction {..} = _cfiFunction
case toList _cffInputArgs of
[] ->
pure $
(fromComputedField cfName,) $
JoinComputedField $
ScalarComputedField
_cfiXComputedFieldInfo
_cfiName
_cffName
_cffComputedFieldImplicitArgs
scalarType
_ -> Nothing
in M.union columnFields computedFields
rawRemoteRelationshipInfos <-
buildInfoMapPreservingMetadata
(_rrName . (^. _3))
server: fix the nullability of object relationships (fix hasura/graphql-engine#7201) When adding object relationships, we set the nullability of the generated GraphQL field based on whether the database backend enforces that the referenced data always exists. For manual relationships (corresponding to `manual_configuration`), the database backend is unaware of any relationship between data, and hence such fields are always set to be nullable. For relationships generated from foreign key constraints (corresponding to `foreign_key_constraint_on`), we distinguish between two cases: 1. The "forward" object relationship from a referencing table (i.e. which has the foreign key constraint) to a referenced table. This should be set to be non-nullable when all referencing columns are non-nullable. But in fact, it used to set it to be non-nullable if *any* referencing column is non-nullable, which is only correct in Postgres when `MATCH FULL` is set (a flag we don't consider). This fixes that by changing a boolean conjunction to a disjunction. 2. The "reverse" object relationship from a referenced table to a referencing table which has the foreign key constraint. This should always be set to be nullable. But in fact, it used to always be set to non-nullable, as was reported in hasura/graphql-engine#7201. This fixes that. Moreover, we have moved the computation of the nullability from `Hasura.RQL.DDL.Relationship` to `Hasura.GraphQL.Schema.Select`: this nullability used to be passed through the `riIsNullable` field of `RelInfo`, but for array relationships this information is not actually used, and moreover the remaining fields of `RelInfo` are already enough to deduce the nullability. This also adds regression tests for both (1) and (2) above. https://github.com/hasura/graphql-engine-mono/pull/2159 GitOrigin-RevId: 617f12765614f49746d18d3368f41dfae2f3e6ca
2021-08-26 18:26:43 +03:00
(mkRemoteRelationshipMetadataObject @b)
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
( proc ((a, b, c), d) -> do
o <- interpretWriter -< buildRemoteRelationship a b c d
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
arrM liftEither -< o
)
server: fix the nullability of object relationships (fix hasura/graphql-engine#7201) When adding object relationships, we set the nullability of the generated GraphQL field based on whether the database backend enforces that the referenced data always exists. For manual relationships (corresponding to `manual_configuration`), the database backend is unaware of any relationship between data, and hence such fields are always set to be nullable. For relationships generated from foreign key constraints (corresponding to `foreign_key_constraint_on`), we distinguish between two cases: 1. The "forward" object relationship from a referencing table (i.e. which has the foreign key constraint) to a referenced table. This should be set to be non-nullable when all referencing columns are non-nullable. But in fact, it used to set it to be non-nullable if *any* referencing column is non-nullable, which is only correct in Postgres when `MATCH FULL` is set (a flag we don't consider). This fixes that by changing a boolean conjunction to a disjunction. 2. The "reverse" object relationship from a referenced table to a referencing table which has the foreign key constraint. This should always be set to be nullable. But in fact, it used to always be set to non-nullable, as was reported in hasura/graphql-engine#7201. This fixes that. Moreover, we have moved the computation of the nullability from `Hasura.RQL.DDL.Relationship` to `Hasura.GraphQL.Schema.Select`: this nullability used to be passed through the `riIsNullable` field of `RelInfo`, but for array relationships this information is not actually used, and moreover the remaining fields of `RelInfo` are already enough to deduce the nullability. This also adds regression tests for both (1) and (2) above. https://github.com/hasura/graphql-engine-mono/pull/2159 GitOrigin-RevId: 617f12765614f49746d18d3368f41dfae2f3e6ca
2021-08-26 18:26:43 +03:00
-<
((allSources, lhsJoinFields, remoteSchemaMap), map (source,_nctiTable,) _nctiRemoteRelationships)
let relationshipFields = mapKeys fromRel relationshipInfos
computedFieldFields = mapKeys fromComputedField computedFieldInfos
remoteRelationshipFields = mapKeys fromRemoteRelationship rawRemoteRelationshipInfos
-- First, check for conflicts between non-column fields, since we can raise a better error
-- message in terms of the two metadata objects that define them.
(align relationshipFields computedFieldFields >- returnA)
>-> (| Inc.keyed (\fieldName fields -> (fieldName, fields) >- noFieldConflicts FIRelationship FIComputedField) |)
-- Second, align with remote relationship fields
>-> (\fields -> align (catMaybes fields) remoteRelationshipFields >- returnA)
>-> (| Inc.keyed (\fieldName fields -> (fieldName, fields) >- noFieldConflicts id FIRemoteRelationship) |)
-- Next, check for conflicts with custom field names. This is easiest to do before merging with
-- the column info itself because we have access to the information separately, and custom field
-- names are not currently stored as a separate map (but maybe should be!).
>-> (\fields -> (columns, catMaybes fields) >- noCustomFieldConflicts)
-- Finally, check for conflicts with the columns themselves.
>-> (\fields -> align columns (catMaybes fields) >- returnA)
>-> (| Inc.keyed (\_ fields -> fields >- noColumnConflicts) |)
where
noFieldConflicts this that = proc (fieldName, fields) -> case fields of
This (thisField, metadata) -> returnA -< Just (this thisField, metadata)
That (thatField, metadata) -> returnA -< Just (that thatField, metadata)
These (_, thisMetadata) (_, thatMetadata) -> do
tellA
-<
Seq.singleton $
CIInconsistency $
ConflictingObjects
("conflicting definitions for field " <>> fieldName)
[thisMetadata, thatMetadata]
returnA -< Nothing
noCustomFieldConflicts = proc (columns, nonColumnFields) -> do
let columnsByGQLName = mapFromL ciName $ M.elems columns
(|
Inc.keyed
( \_ (fieldInfo, metadata) ->
(|
withRecordInconsistency
( do
(|
traverseA_
( \fieldGQLName -> case M.lookup fieldGQLName columnsByGQLName of
-- Only raise an error if the GQL name isnt the same as the Postgres column name.
-- If they are the same, `noColumnConflicts` will catch it, and it will produce a
-- more useful error message.
Just columnInfo
| toTxt (ciColumn columnInfo) /= G.unName fieldGQLName ->
throwA
-<
err400 AlreadyExists $
"field definition conflicts with custom field name for postgres column "
<>> ciColumn columnInfo
_ -> returnA -< ()
)
|) (fieldInfoGraphQLNames fieldInfo)
returnA -< (fieldInfo, metadata)
)
|) metadata
)
|) nonColumnFields
noColumnConflicts = proc fields -> case fields of
This columnInfo -> returnA -< FIColumn columnInfo
That (fieldInfo, _) -> returnA -< fieldInfo
These columnInfo (_, fieldMetadata) -> do
recordInconsistency -< ((Nothing, fieldMetadata), "field definition conflicts with postgres column")
returnA -< FIColumn columnInfo
mkRelationshipMetadataObject ::
forall b a.
(ToJSON a, Backend b) =>
RelType ->
(SourceName, TableName b, RelDef a) ->
MetadataObject
mkRelationshipMetadataObject relType (source, table, relDef) =
let objectId =
MOSourceObjId source $
AB.mkAnyBackend $
SMOTableObj @b table $
MTORel (_rdName relDef) relType
in MetadataObject objectId $ toJSON $ WithTable @b source table relDef
buildObjectRelationship ::
( ArrowChoice arr,
ArrowWriter (Seq CollectedInfo) arr,
Backend b
) =>
( HashMap (TableName b) (HashSet (ForeignKey b)),
( SourceName,
TableName b,
ObjRelDef b
)
)
`arr` Maybe (RelInfo b)
server: fix the nullability of object relationships (fix hasura/graphql-engine#7201) When adding object relationships, we set the nullability of the generated GraphQL field based on whether the database backend enforces that the referenced data always exists. For manual relationships (corresponding to `manual_configuration`), the database backend is unaware of any relationship between data, and hence such fields are always set to be nullable. For relationships generated from foreign key constraints (corresponding to `foreign_key_constraint_on`), we distinguish between two cases: 1. The "forward" object relationship from a referencing table (i.e. which has the foreign key constraint) to a referenced table. This should be set to be non-nullable when all referencing columns are non-nullable. But in fact, it used to set it to be non-nullable if *any* referencing column is non-nullable, which is only correct in Postgres when `MATCH FULL` is set (a flag we don't consider). This fixes that by changing a boolean conjunction to a disjunction. 2. The "reverse" object relationship from a referenced table to a referencing table which has the foreign key constraint. This should always be set to be nullable. But in fact, it used to always be set to non-nullable, as was reported in hasura/graphql-engine#7201. This fixes that. Moreover, we have moved the computation of the nullability from `Hasura.RQL.DDL.Relationship` to `Hasura.GraphQL.Schema.Select`: this nullability used to be passed through the `riIsNullable` field of `RelInfo`, but for array relationships this information is not actually used, and moreover the remaining fields of `RelInfo` are already enough to deduce the nullability. This also adds regression tests for both (1) and (2) above. https://github.com/hasura/graphql-engine-mono/pull/2159 GitOrigin-RevId: 617f12765614f49746d18d3368f41dfae2f3e6ca
2021-08-26 18:26:43 +03:00
buildObjectRelationship = proc (fkeysMap, (source, table, relDef)) -> do
let buildRelInfo def = objRelP2Setup source table fkeysMap def
interpretWriter -< buildRelationship source table buildRelInfo ObjRel relDef
buildArrayRelationship ::
( ArrowChoice arr,
ArrowWriter (Seq CollectedInfo) arr,
Backend b
) =>
( HashMap (TableName b) (HashSet (ForeignKey b)),
( SourceName,
TableName b,
ArrRelDef b
)
)
`arr` Maybe (RelInfo b)
buildArrayRelationship = proc (fkeysMap, (source, table, relDef)) -> do
let buildRelInfo def = arrRelP2Setup fkeysMap source table def
interpretWriter -< buildRelationship source table buildRelInfo ArrRel relDef
buildRelationship ::
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
forall m b a.
( MonadWriter (Seq CollectedInfo) m,
ToJSON a,
Backend b
) =>
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
SourceName ->
TableName b ->
(RelDef a -> Either QErr (RelInfo b, [SchemaDependency])) ->
RelType ->
RelDef a ->
m (Maybe (RelInfo b))
buildRelationship source table buildRelInfo relType relDef = do
let relName = _rdName relDef
metadataObject = mkRelationshipMetadataObject @b relType (source, table, relDef)
schemaObject =
SOSourceObj source $
AB.mkAnyBackend $
SOITableObj @b table $
TORel relName
addRelationshipContext e = "in relationship " <> relName <<> ": " <> e
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
withRecordInconsistencyM metadataObject $ do
modifyErr (addTableContext @b table . addRelationshipContext) $ do
(info, dependencies) <- liftEither $ buildRelInfo relDef
recordDependenciesM metadataObject schemaObject dependencies
return info
mkComputedFieldMetadataObject ::
forall b.
(Backend b) =>
(SourceName, TableName b, ComputedFieldMetadata b) ->
MetadataObject
mkComputedFieldMetadataObject (source, table, ComputedFieldMetadata {..}) =
let objectId =
MOSourceObjId source $
AB.mkAnyBackend $
SMOTableObj @b table $
MTOComputedField _cfmName
definition = AddComputedField @b source table _cfmName _cfmDefinition _cfmComment
in MetadataObject objectId (toJSON definition)
buildComputedField ::
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
forall b m.
( MonadWriter (Seq CollectedInfo) m,
BackendMetadata b
) =>
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
HashSet (TableName b) ->
HashSet (Column b) ->
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
SourceName ->
DBFunctionsMetadata b ->
TableName b ->
ComputedFieldMetadata b ->
m (Either QErr (Maybe (ComputedFieldInfo b)))
buildComputedField trackedTableNames tableColumns source pgFunctions table cf@ComputedFieldMetadata {..} = runExceptT do
let addComputedFieldContext e = "in computed field " <> _cfmName <<> ": " <> e
function = computedFieldFunction @b _cfmDefinition
funcDefs = fromMaybe [] $ M.lookup function pgFunctions
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
withRecordInconsistencyM (mkComputedFieldMetadataObject (source, table, cf)) $
modifyErr (addTableContext @b table . addComputedFieldContext) $ do
rawfi <- handleMultipleFunctions @b (computedFieldFunction @b _cfmDefinition) funcDefs
buildComputedFieldInfo trackedTableNames table tableColumns _cfmName _cfmDefinition rawfi _cfmComment
mkRemoteRelationshipMetadataObject ::
forall b.
Backend b =>
(SourceName, TableName b, RemoteRelationship) ->
MetadataObject
mkRemoteRelationshipMetadataObject (source, table, RemoteRelationship {..}) =
let objectId =
MOSourceObjId source $
AB.mkAnyBackend $
SMOTableObj @b table $
MTORemoteRelationship _rrName
in MetadataObject objectId $
toJSON $
CreateFromSourceRelationship @b source table _rrName _rrDefinition
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
-- | This is a "thin" wrapper around 'buildRemoteFieldInfo', which only knows
-- how to construct dependencies on the RHS of the join condition, so the
-- dependencies on the remote relationship on the LHS entity are computed here
buildRemoteRelationship ::
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
forall b m.
( MonadWriter (Seq CollectedInfo) m,
BackendMetadata b
) =>
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
HashMap SourceName (AB.AnyBackend PartiallyResolvedSource) ->
M.HashMap FieldName (DBJoinField b) ->
scaffolding for remote-schemas module The main aim of the PR is: 1. To set up a module structure for 'remote-schemas' package. 2. Move parts by the remote schema codebase into the new module structure to validate it. ## Notes to the reviewer Why a PR with large-ish diff? 1. We've been making progress on the MM project but we don't yet know long it is going to take us to get to the first milestone. To understand this better, we need to figure out the unknowns as soon as possible. Hence I've taken a stab at the first two items in the [end-state](https://gist.github.com/0x777/ca2bdc4284d21c3eec153b51dea255c9) document to figure out the unknowns. Unsurprisingly, there are a bunch of issues that we haven't discussed earlier. These are documented in the 'open questions' section. 1. The diff is large but that is only code moved around and I've added a section that documents how things are moved. In addition, there are fair number of PR comments to help with the review process. ## Changes in the PR ### Module structure Sets up the module structure as follows: ``` Hasura/ RemoteSchema/ Metadata/ Types.hs SchemaCache/ Types.hs Permission.hs RemoteRelationship.hs Build.hs MetadataAPI/ Types.hs Execute.hs ``` ### 1. Types representing metadata are moved Types that capture metadata information (currently scattered across several RQL modules) are moved into `Hasura.RemoteSchema.Metadata.Types`. - This new module only depends on very 'core' modules such as `Hasura.Session` for the notion of roles and `Hasura.Incremental` for `Cacheable` typeclass. - The requirement on database modules is avoided by generalizing the remote schemas metadata to accept an arbitrary 'r' for a remote relationship definition. ### 2. SchemaCache related types and build logic have been moved Types that represent remote schemas information in SchemaCache are moved into `Hasura.RemoteSchema.SchemaCache.Types`. Similar to `H.RS.Metadata.Types`, this module depends on 'core' modules except for `Hasura.GraphQL.Parser.Variable`. It has something to do with remote relationships but I haven't spent time looking into it. The validation of 'remote relationships to remote schema' is also something that needs to be looked at. Rips out the logic that builds remote schema's SchemaCache information from the monolithic `buildSchemaCacheRule` and moves it into `Hasura.RemoteSchema.SchemaCache.Build`. Further, the `.SchemaCache.Permission` and `.SchemaCache.RemoteRelationship` have been created from existing modules that capture schema cache building logic for those two components. This was a fair amount of work. On main, currently remote schema's SchemaCache information is built in two phases - in the first phase, 'permissions' and 'remote relationships' are ignored and in the second phase they are filled in. While remote relationships can only be resolved after partially resolving sources and other remote schemas, the same isn't true for permissions. Further, most of the work that is done to resolve remote relationships can be moved to the first phase so that the second phase can be a very simple traversal. This is the approach that was taken - resolve permissions and as much as remote relationships information in the first phase. ### 3. Metadata APIs related types and build logic have been moved The types that represent remote schema related metadata APIs and the execution logic have been moved to `Hasura.RemoteSchema.MetadataAPI.Types` and `.Execute` modules respectively. ## Open questions: 1. `Hasura.RemoteSchema.Metadata.Types` is so called because I was hoping that all of the metadata related APIs of remote schema can be brought in at `Hasura.RemoteSchema.Metadata.API`. However, as metadata APIs depended on functions from `SchemaCache` module (see [1](https://github.com/hasura/graphql-engine-mono/blob/ceba6d62264603ee5d279814677b29bcc43ecaea/server/src-lib/Hasura/RQL/DDL/RemoteSchema.hs#L55) and [2](https://github.com/hasura/graphql-engine-mono/blob/ceba6d62264603ee5d279814677b29bcc43ecaea/server/src-lib/Hasura/RQL/DDL/RemoteSchema.hs#L91), it made more sense to create a separate top-level module for `MetadataAPI`s. Maybe we can just have `Hasura.RemoteSchema.Metadata` and get rid of the extra nesting or have `Hasura.RemoteSchema.Metadata.{Core,Permission,RemoteRelationship}` if we want to break them down further. 1. `buildRemoteSchemas` in `H.RS.SchemaCache.Build` has the following type: ```haskell buildRemoteSchemas :: ( ArrowChoice arr, Inc.ArrowDistribute arr, ArrowWriter (Seq CollectedInfo) arr, Inc.ArrowCache m arr, MonadIO m, HasHttpManagerM m, Inc.Cacheable remoteRelationshipDefinition, ToJSON remoteRelationshipDefinition, MonadError QErr m ) => Env.Environment -> ( (Inc.Dependency (HashMap RemoteSchemaName Inc.InvalidationKey), OrderedRoles), [RemoteSchemaMetadataG remoteRelationshipDefinition] ) `arr` HashMap RemoteSchemaName (PartiallyResolvedRemoteSchemaCtxG remoteRelationshipDefinition, MetadataObject) ``` Note the dependence on `CollectedInfo` which is defined as ```haskell data CollectedInfo = CIInconsistency InconsistentMetadata | CIDependency MetadataObject -- ^ for error reporting on missing dependencies SchemaObjId SchemaDependency deriving (Eq) ``` this pretty much means that remote schemas is dependent on types from databases, actions, .... How do we fix this? Maybe introduce a typeclass such as `ArrowCollectRemoteSchemaDependencies` which is defined in `Hasura.RemoteSchema` and then implemented in graphql-engine? 1. The dependency on `buildSchemaCacheFor` in `.MetadataAPI.Execute` which has the following signature: ```haskell buildSchemaCacheFor :: (QErrM m, CacheRWM m, MetadataM m) => MetadataObjId -> MetadataModifier -> ``` This can be easily resolved if we restrict what the metadata APIs are allowed to do. Currently, they operate in an unfettered access to modify SchemaCache (the `CacheRWM` constraint): ```haskell runAddRemoteSchema :: ( QErrM m, CacheRWM m, MonadIO m, HasHttpManagerM m, MetadataM m, Tracing.MonadTrace m ) => Env.Environment -> AddRemoteSchemaQuery -> m EncJSON ``` This should instead be changed to restrict remote schema APIs to only modify remote schema metadata (but has access to the remote schemas part of the schema cache), this dependency is completely removed. ```haskell runAddRemoteSchema :: ( QErrM m, MonadIO m, HasHttpManagerM m, MonadReader RemoteSchemasSchemaCache m, MonadState RemoteSchemaMetadata m, Tracing.MonadTrace m ) => Env.Environment -> AddRemoteSchemaQuery -> m RemoteSchemeMetadataObjId ``` The idea is that the core graphql-engine would call these functions and then call `buildSchemaCacheFor`. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/6291 GitOrigin-RevId: 51357148c6404afe70219afa71bd1d59bdf4ffc6
2022-10-21 06:13:07 +03:00
PartiallyResolvedRemoteSchemaMap ->
Avoid `Arrows` by interpreting monads TL;DR --- We go from this: ```haskell (| withRecordInconsistency ( (| modifyErrA ( do (info, dependencies) <- liftEitherA -< buildRelInfo relDef recordDependencies -< (metadataObject, schemaObject, dependencies) returnA -< info ) |) (addTableContext @b table . addRelationshipContext) ) |) metadataObject ``` to this: ```haskell withRecordInconsistencyM metadataObject $ do modifyErr (addTableContext @b table . addRelationshipContext) $ do (info, dependencies) <- liftEither $ buildRelInfo relDef recordDependenciesM metadataObject schemaObject dependencies return info ``` Background --- We use Haskell's `Arrows` language extension to gain some syntactic sugar when working with `Arrow`s. `Arrow`s are a programming abstraction comparable to `Monad`s. Unfortunately the syntactic sugar provided by this language extension is not very sweet. This PR shows how we can sometimes avoid using `Arrow`s altogether, without loss of functionality or correctness. It is a demo of a technique that can be used to cut down the amount of `Arrows`-based code in our codebase by about half. Approach --- Although _in general_ not every `Monad` is an `Arrow`, specific `Arrow` instantiations are exactly as powerful as their `Monad` equivalents. Otherwise they wouldn't be very equivalent, would they? Just like `liftEither` interprets the `Either e` monad into an arbitrary monad implementing `MonadError e`, we add `interpA` which interprets certain concrete monads such as `Writer w` into specific arrows, e.g. ones satisfying `ArrowWriter w`. This means that the part of the code that only uses such interpretable effects can be written _monadically_, and then used in _arrow_ constructions down the line. This approach cannot be used for arrow effects which do not have a monadic equivalent. In our codebase, the only instance of this is `ArrowCache m`, implemented by the `Rule m` arrow. So code written with `ArrowCache m` in the context cannot be rewritten monadically using this technique. See also --- - #1827 - #2210 PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3543 Co-authored-by: jkachmar <8461423+jkachmar@users.noreply.github.com> GitOrigin-RevId: eb79619c95f7a571bce99bc144ce42ee65d08505
2022-02-22 21:08:54 +03:00
(SourceName, TableName b, RemoteRelationship) ->
m (Either QErr (Maybe (RemoteFieldInfo (DBJoinField b))))
buildRemoteRelationship allSources allColumns remoteSchemaMap (source, table, rr@RemoteRelationship {..}) = runExceptT $ do
let metadataObject = mkRemoteRelationshipMetadataObject @b (source, table, rr)
schemaObj =
SOSourceObj source $
AB.mkAnyBackend $
SOITableObj @b table $
TORemoteRel _rrName
addRemoteRelationshipContext e = "in remote relationship" <> _rrName <<> ": " <> e
withRecordInconsistencyM metadataObject $
modifyErr (addTableContext @b table . addRemoteRelationshipContext) $ do
(remoteField, rhsDependencies) <-
buildRemoteFieldInfo (tableNameToLHSIdentifier @b table) allColumns rr allSources remoteSchemaMap
let lhsDependencies =
-- a direct dependency on the table on which this is defined
SchemaDependency (SOSourceObj source $ AB.mkAnyBackend $ SOITable @b table) DRTable
-- the relationship is also dependent on all the lhs
-- columns that are used in the join condition
: flip map (M.elems $ _rfiLHS remoteField) \case
JoinColumn column _ ->
-- TODO: shouldn't this be DRColumn??
mkColDep @b DRRemoteRelationship source table column
JoinComputedField computedFieldInfo ->
mkComputedFieldDep @b DRRemoteRelationship source table $ _scfName computedFieldInfo
-- Here is the essence of the function: construct dependencies on the RHS
-- of the join condition.
recordDependenciesM metadataObject schemaObj (lhsDependencies <> rhsDependencies)
return remoteField