graphql-engine/server/src-lib/Hasura/GraphQL/Schema/Build.hs

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

373 lines
18 KiB
Haskell
Raw Normal View History

-- | This module provides building blocks for the GraphQL Schema that the
-- GraphQL Engine presents.
--
-- The functions defined here are used to serve as default implementations for
-- their namesakes in the 'BackendSchema' type class.
--
-- When, for some backend, you want to implement a new feature that manifests
-- itself visibly in the schema (e.g., if you're developing support for update
-- mutations), this module is likely where your efforts should start.
--
-- Using these functions help us present a consistent GraphQL schema across
-- different backends.
--
-- There is a bit of tension however, as sometimes we intentionally do want the
-- GraphQL Schema relating to some backend to be different in some way.
--
-- It could be that a backend only has limited support for some common feature,
-- or, more interestingly, that some backend just does things differently (c.f.
-- MSSQL's @MERGE@ statement with PostgreSQL's @INSERT .. ON CONFLICT@, which
-- are similar enough that we want to use the same overall upsert schema but
-- different enough that we want to use different field names)
--
-- When you want to implement new schema for a backend, there is overall three
-- different ways do deal with this tension:
--
-- 1. You can duplicate existing code and implement the new behavior in the
-- duplicate.
-- 2. You can infuse the new behavior into existing code and switch dynamically
-- at runtime (or via type class instance dispatch, which is the same
-- for our purposes)
-- 3. You can refactor the existing building blocks and compose them differently
-- at use sites to get the desired behavior nuances.
--
-- Of these three, steps 1. and 2. are by far the easiest to execute, while 3.
-- requires some critical thought. However, both 1. and 2. produce legacy code
-- that is difficult to maintain and understand.
--
-- As a guideline, if you find yourself wanting add new behavior to some of
-- these functions it's very likely that you should consider refactoring them
-- instead, thus shifting the responsibility deciding on the correct behavior to
-- use sites.
--
-- It an ongoing effort to adapt and refactor these building blocks such that
-- they have the sizes and shapes that result in the most elegant uses of them
-- that we can manage.
module Hasura.GraphQL.Schema.Build
( buildTableDeleteMutationFields,
buildTableInsertMutationFields,
buildTableQueryAndSubscriptionFields,
buildTableStreamingSubscriptionFields,
buildTableUpdateMutationFields,
setFieldNameCase,
buildFieldDescription,
)
where
import Data.Has (getter)
import Data.Text.Casing qualified as C
import Data.Text.Extended
import Hasura.GraphQL.ApolloFederation
import Hasura.GraphQL.Schema.Backend (BackendTableSelectSchema (..), MonadBuildSchema)
import Hasura.GraphQL.Schema.BoolExp (AggregationPredicatesSchema)
import Hasura.GraphQL.Schema.Common
import Hasura.GraphQL.Schema.Mutation
import Hasura.GraphQL.Schema.NamingCase
import Hasura.GraphQL.Schema.Options qualified as Options
server: Metadata origin for definitions (type parameter version v2) The code that builds the GraphQL schema, and `buildGQLContext` in particular, is partial: not every value of `(ServerConfigCtx, GraphQLQueryType, SourceCache, HashMap RemoteSchemaName (RemoteSchemaCtx, MetadataObject), ActionCache, AnnotatedCustomTypes)` results in a valid GraphQL schema. When it fails, we want to be able to return better error messages than we currently do. The key thing that is missing is a way to trace back GraphQL type information to their origin from the Hasura metadata. Currently, we have a number of correctness checks of our GraphQL schema. But these correctness checks only have access to pure GraphQL type information, and hence can only report errors in terms of that. Possibly the worst is the "conflicting definitions" error, which, in practice, can only be debugged by Hasura engineers. This is terrible DX for customers. This PR allows us to print better error messages, by adding a field to the `Definition` type that traces the GraphQL type to its origin in the metadata. So the idea is simple: just add `MetadataObjId`, or `Maybe` that, or some other sum type of that, to `Definition`. However, we want to avoid having to import a `Hasura.RQL` module from `Hasura.GraphQL.Parser`. So we instead define this additional field of `Definition` through a new type parameter, which is threaded through in `Hasura.GraphQL.Parser`. We then define type synonyms in `Hasura.GraphQL.Schema.Parser` that fill in this type parameter, so that it is not visible for the majority of the codebase. The idea of associating metadata information to `Definition`s really comes to fruition when combined with hasura/graphql-engine-mono#4517. Their combination would allow us to use the API of fatal errors (just like the current `MonadError QErr`) to report _inconsistencies_ in the metadata. Such inconsistencies are then _automatically_ ignored. So no ad-hoc decisions need to be made on how to cut out inconsistent metadata from the GraphQL schema. This will allow us to report much better errors, as well as improve the likelihood of a successful HGE startup. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/4770 Co-authored-by: Samir Talwar <47582+SamirTalwar@users.noreply.github.com> GitOrigin-RevId: 728402b0cae83ae8e83463a826ceeb609001acae
2022-06-28 18:52:26 +03:00
import Hasura.GraphQL.Schema.Parser hiding (EnumValueInfo, field)
import Hasura.GraphQL.Schema.Select
import Hasura.GraphQL.Schema.SubscriptionStream (selectStreamTable)
import Hasura.GraphQL.Schema.Table (getTableIdentifierName, tableSelectPermissions)
import Hasura.GraphQL.Schema.Typename
import Hasura.GraphQL.Schema.Update (updateTable, updateTableByPk)
import Hasura.Prelude
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.IR
import Hasura.RQL.Types.Backend
import Hasura.RQL.Types.Common
import Hasura.RQL.Types.Permission
import Hasura.RQL.Types.SchemaCache
2022-05-27 20:21:22 +03:00
import Hasura.RQL.Types.Source
import Hasura.RQL.Types.SourceCustomization
import Hasura.RQL.Types.Table
import Language.GraphQL.Draft.Syntax qualified as G
-- | Builds field name with proper case. Please note that this is a pure
-- function as all the validation has already been done while preparing
-- @GQLNameIdentifier@.
setFieldNameCase ::
NamingCase ->
TableInfo b ->
CustomRootField ->
(C.GQLNameIdentifier -> C.GQLNameIdentifier) ->
C.GQLNameIdentifier ->
G.Name
setFieldNameCase tCase tInfo crf getFieldName tableName =
(applyFieldNameCaseIdentifier tCase fieldIdentifier)
where
tccName = fmap C.fromCustomName . _tcCustomName . _tciCustomConfig . _tiCoreInfo $ tInfo
crfName = fmap C.fromCustomName (_crfName crf)
fieldIdentifier = fromMaybe (getFieldName (fromMaybe tableName tccName)) crfName
-- | buildTableQueryAndSubscriptionFields builds the field parsers of a table.
-- It returns a tuple with array of field parsers that correspond to the field
-- parsers of the query root and the field parsers of the subscription root
buildTableQueryAndSubscriptionFields ::
forall b r m n.
( MonadBuildSchema b r m n,
AggregationPredicatesSchema b,
BackendTableSelectSchema b
) =>
MkRootFieldName ->
TableName b ->
TableInfo b ->
C.GQLNameIdentifier ->
SchemaT
r
m
( [FieldParser n (QueryDB b (RemoteRelationshipField UnpreparedValue) (UnpreparedValue b))],
[FieldParser n (QueryDB b (RemoteRelationshipField UnpreparedValue) (UnpreparedValue b))],
Maybe (G.Name, Parser 'Output n (ApolloFederationParserFunction n))
)
buildTableQueryAndSubscriptionFields mkRootFieldName tableName tableInfo gqlName = do
sourceInfo :: SourceInfo b <- asks getter
Move RoleName into SchemaContext. ### Description I am not 100% sure about this PR; while I think the code is better this way, I'm willing to be convinced otherwise. In short, this PR moves the `RoleName` field into the `SchemaContext`, instead of being a nebulous `Has RoleName` constraint on the reader monad. The major upside of this is that it makes it an explicit named field, rather than something that must be given as part of a tuple of arguments when calling `runReader`. However, the downside is that it breaks the helper permissions functions of `Schema.Table`, which relied on `Has RoleName r`. This PR makes the choice of passing the role name explicitly to all of those functions, which in turn means first explicitly fetching the role name in a lot of places. It makes it more explicit when a schema building block relies on the role name, but is a bit verbose... ### Alternatives Some alternatives worth considering: - attempting something like `Has context r, Has RoleName context`, which would allow them to be independent from the context but still fetch the role name from the reader, but might require type annotations to not be ambiguous - keeping the permission functions the same, with `Has RoleName r`, and introducing a bunch of newtypes instead of using tuples to explicitly implement all the required `Has` instances - changing the permission functions to `Has SchemaContext r`, since they are functions used only to build the schema, and therefore may be allowed to be tied to the context. What do y'all think? PR-URL: https://github.com/hasura/graphql-engine-mono/pull/5073 GitOrigin-RevId: 8fd09fafb54905a4d115ef30842d35da0c3db5d2
2022-07-29 18:37:09 +03:00
roleName <- retrieve scRole
let customization = _siCustomization sourceInfo
tCase = _rscNamingConvention customization
mkTypename = runMkTypename $ _rscTypeNames customization
-- select table
selectName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfSelect mkSelectField gqlName
-- select table by pk
selectPKName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfSelectByPk mkSelectByPkField gqlName
-- select table aggregate
selectAggName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfSelectAggregate mkSelectAggregateField gqlName
selectTableParser <- optionalFieldParser QDBMultipleRows $ selectTable tableInfo selectName selectDesc
selectTableByPkParser <- optionalFieldParser QDBSingleRow $ selectTableByPk tableInfo selectPKName selectPKDesc
selectTableAggregateParser <- optionalFieldParser QDBAggregation $ selectTableAggregate tableInfo selectAggName selectAggDesc
Move RoleName into SchemaContext. ### Description I am not 100% sure about this PR; while I think the code is better this way, I'm willing to be convinced otherwise. In short, this PR moves the `RoleName` field into the `SchemaContext`, instead of being a nebulous `Has RoleName` constraint on the reader monad. The major upside of this is that it makes it an explicit named field, rather than something that must be given as part of a tuple of arguments when calling `runReader`. However, the downside is that it breaks the helper permissions functions of `Schema.Table`, which relied on `Has RoleName r`. This PR makes the choice of passing the role name explicitly to all of those functions, which in turn means first explicitly fetching the role name in a lot of places. It makes it more explicit when a schema building block relies on the role name, but is a bit verbose... ### Alternatives Some alternatives worth considering: - attempting something like `Has context r, Has RoleName context`, which would allow them to be independent from the context but still fetch the role name from the reader, but might require type annotations to not be ambiguous - keeping the permission functions the same, with `Has RoleName r`, and introducing a bunch of newtypes instead of using tuples to explicitly implement all the required `Has` instances - changing the permission functions to `Has SchemaContext r`, since they are functions used only to build the schema, and therefore may be allowed to be tied to the context. What do y'all think? PR-URL: https://github.com/hasura/graphql-engine-mono/pull/5073 GitOrigin-RevId: 8fd09fafb54905a4d115ef30842d35da0c3db5d2
2022-07-29 18:37:09 +03:00
case tableSelectPermissions roleName tableInfo of
-- No select permission found for the current role, so
-- no root fields will be accessible to the role
Nothing -> pure (mempty, mempty, Nothing)
-- Filter the root fields which have been enabled
Just SelPermInfo {..} -> do
selectStreamParser <-
if (isRootFieldAllowed SRFTSelectStream spiAllowedSubscriptionRootFields)
then buildTableStreamingSubscriptionFields mkRootFieldName tableName tableInfo gqlName
else pure mempty
let (querySelectTableParser, subscriptionSelectTableParser) =
getQueryAndSubscriptionRootFields
selectTableParser
(isRootFieldAllowed QRFTSelect spiAllowedQueryRootFields)
(isRootFieldAllowed SRFTSelect spiAllowedSubscriptionRootFields)
(querySelectTableByPkParser, subscriptionSelectTableByPkParser) =
getQueryAndSubscriptionRootFields
selectTableByPkParser
(isRootFieldAllowed QRFTSelectByPk spiAllowedQueryRootFields)
(isRootFieldAllowed SRFTSelectByPk spiAllowedSubscriptionRootFields)
(querySelectTableAggParser, subscriptionSelectTableAggParser) =
getQueryAndSubscriptionRootFields
selectTableAggregateParser
(isRootFieldAllowed QRFTSelectAggregate spiAllowedQueryRootFields)
(isRootFieldAllowed SRFTSelectAggregate spiAllowedSubscriptionRootFields)
queryRootFields = catMaybes [querySelectTableParser, querySelectTableByPkParser, querySelectTableAggParser]
subscriptionRootFields =
selectStreamParser
<> catMaybes [subscriptionSelectTableParser, subscriptionSelectTableByPkParser, subscriptionSelectTableAggParser]
-- This parser is for generating apollo federation field _entities
apolloFedTableParser <- runMaybeT do
guard $ isApolloFedV1enabled (_tciApolloFederationConfig (_tiCoreInfo tableInfo))
tableSelSet <- MaybeT $ tableSelectionSet tableInfo
Move RoleName into SchemaContext. ### Description I am not 100% sure about this PR; while I think the code is better this way, I'm willing to be convinced otherwise. In short, this PR moves the `RoleName` field into the `SchemaContext`, instead of being a nebulous `Has RoleName` constraint on the reader monad. The major upside of this is that it makes it an explicit named field, rather than something that must be given as part of a tuple of arguments when calling `runReader`. However, the downside is that it breaks the helper permissions functions of `Schema.Table`, which relied on `Has RoleName r`. This PR makes the choice of passing the role name explicitly to all of those functions, which in turn means first explicitly fetching the role name in a lot of places. It makes it more explicit when a schema building block relies on the role name, but is a bit verbose... ### Alternatives Some alternatives worth considering: - attempting something like `Has context r, Has RoleName context`, which would allow them to be independent from the context but still fetch the role name from the reader, but might require type annotations to not be ambiguous - keeping the permission functions the same, with `Has RoleName r`, and introducing a bunch of newtypes instead of using tuples to explicitly implement all the required `Has` instances - changing the permission functions to `Has SchemaContext r`, since they are functions used only to build the schema, and therefore may be allowed to be tied to the context. What do y'all think? PR-URL: https://github.com/hasura/graphql-engine-mono/pull/5073 GitOrigin-RevId: 8fd09fafb54905a4d115ef30842d35da0c3db5d2
2022-07-29 18:37:09 +03:00
selectPerm <- hoistMaybe $ tableSelectPermissions roleName tableInfo
stringifyNumbers <- retrieve Options.soStringifyNumbers
primaryKeys <- hoistMaybe $ fmap _pkColumns . _tciPrimaryKey . _tiCoreInfo $ tableInfo
let tableSelPerm = tablePermissionsInfo selectPerm
tableGQLName <- getTableIdentifierName tableInfo
let objectTypename = mkTypename $ applyTypeNameCaseIdentifier tCase $ mkTableTypeName $ tableGQLName
pure $ (objectTypename, convertToApolloFedParserFunc sourceInfo tableInfo tableSelPerm stringifyNumbers (Just tCase) primaryKeys tableSelSet)
pure (queryRootFields, subscriptionRootFields, apolloFedTableParser)
where
selectDesc = buildFieldDescription defaultSelectDesc $ _crfComment _tcrfSelect
selectPKDesc = buildFieldDescription defaultSelectPKDesc $ _crfComment _tcrfSelectByPk
selectAggDesc = buildFieldDescription defaultSelectAggDesc $ _crfComment _tcrfSelectAggregate
defaultSelectDesc = "fetch data from the table: " <>> tableName
defaultSelectPKDesc = "fetch data from the table: " <> tableName <<> " using primary key columns"
defaultSelectAggDesc = "fetch aggregated fields from the table: " <>> tableName
TableCustomRootFields {..} = _tcCustomRootFields . _tciCustomConfig $ _tiCoreInfo tableInfo
-- This function checks if a root field is allowed to be exposed
-- in the query root and a subscription root and when it is allowed,
-- the parser will be returned.
getQueryAndSubscriptionRootFields parser allowedInQuery allowedInSubscription =
case (allowedInQuery, allowedInSubscription) of
(True, True) -> (parser, parser)
(True, False) -> (parser, Nothing)
(False, True) -> (Nothing, parser)
(False, False) -> (Nothing, Nothing)
buildTableStreamingSubscriptionFields ::
forall b r m n.
( MonadBuildSchema b r m n,
AggregationPredicatesSchema b,
BackendTableSelectSchema b
) =>
MkRootFieldName ->
TableName b ->
TableInfo b ->
C.GQLNameIdentifier ->
SchemaT r m [FieldParser n (QueryDB b (RemoteRelationshipField UnpreparedValue) (UnpreparedValue b))]
buildTableStreamingSubscriptionFields mkRootFieldName tableName tableInfo tableIdentifier = do
-- Check in schema options whether we should include streaming subscription
-- fields
include <- retrieve Options.soIncludeStreamFields
case include of
Options.Don'tIncludeStreamFields -> pure mempty
Options.IncludeStreamFields -> do
sourceInfo :: SourceInfo b <- asks getter
let customization = _siCustomization sourceInfo
tCase = _rscNamingConvention customization
customRootFields = _tcCustomRootFields $ _tciCustomConfig $ _tiCoreInfo tableInfo
selectDesc = Just $ G.Description $ "fetch data from the table in a streaming manner: " <>> tableName
selectStreamName =
runMkRootFieldName mkRootFieldName $
setFieldNameCase tCase tableInfo (_tcrfSelectStream customRootFields) mkSelectStreamField tableIdentifier
catMaybes
<$> sequenceA
[ optionalFieldParser QDBStreamMultipleRows $ selectStreamTable tableInfo selectStreamName selectDesc
]
buildTableInsertMutationFields ::
forall b r m n.
( MonadBuildSchema b r m n,
BackendTableSelectSchema b
) =>
(TableInfo b -> SchemaT r m (InputFieldsParser n (BackendInsert b (UnpreparedValue b)))) ->
MkRootFieldName ->
Role-invariant schema constructors We build the GraphQL schema by combining building blocks such as `tableSelectionSet` and `columnParser`. These building blocks individually build `{InputFields,Field,}Parser` objects. Those object specify the valid GraphQL schema. Since the GraphQL schema is role-dependent, at some point we need to know what fragment of the GraphQL schema a specific role is allowed to access, and this is stored in `{Sel,Upd,Ins,Del}PermInfo` objects. We have passed around these permission objects as function arguments to the schema building blocks since we first started dealing with permissions during the PDV refactor - see hasura/graphql-engine@5168b99e463199b1934d8645bd6cd37eddb64ae1 in hasura/graphql-engine#4111. This means that, for instance, `tableSelectionSet` has as its type: ```haskell tableSelectionSet :: forall b r m n. MonadBuildSchema b r m n => SourceName -> TableInfo b -> SelPermInfo b -> m (Parser 'Output n (AnnotatedFields b)) ``` There are three reasons to change this. 1. We often pass a `Maybe (xPermInfo b)` instead of a proper `xPermInfo b`, and it's not clear what the intended semantics of this is. Some potential improvements on the data types involved are discussed in issue hasura/graphql-engine-mono#3125. 2. In most cases we also already pass a `TableInfo b`, and together with the `MonadRole` that is usually also in scope, this means that we could look up the required permissions regardless: so passing the permissions explicitly undermines the "single source of truth" principle. Breaking this principle also makes the code more difficult to read. 3. We are working towards role-based parsers (see hasura/graphql-engine-mono#2711), where the `{InputFields,Field,}Parser` objects are constructed in a role-invariant way, so that we have a single object that can be used for all roles. In particular, this means that the schema building blocks _need_ to be constructed in a role-invariant way. While this PR doesn't accomplish that, it does reduce the amount of role-specific arguments being passed, thus fixing hasura/graphql-engine-mono#3068. Concretely, this PR simply drops the `xPermInfo b` argument from almost all schema building blocks. Instead these objects are looked up from the `TableInfo b` as-needed. The resulting code is considerably simpler and shorter. One way to interpret this change is as follows. Before this PR, we figured out permissions at the top-level in `Hasura.GraphQL.Schema`, passing down the obtained `xPermInfo` objects as required. After this PR, we have a bottom-up approach where the schema building blocks themselves decide whether they want to be included for a particular role. So this moves some permission logic out of `Hasura.GraphQL.Schema`, which is very complex. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3608 GitOrigin-RevId: 51a744f34ec7d57bc8077667ae7f9cb9c4f6c962
2022-02-17 11:16:20 +03:00
Scenario ->
TableName b ->
TableInfo b ->
C.GQLNameIdentifier ->
SchemaT r m [FieldParser n (AnnotatedInsert b (RemoteRelationshipField UnpreparedValue) (UnpreparedValue b))]
buildTableInsertMutationFields backendInsertAction mkRootFieldName scenario tableName tableInfo gqlName = do
sourceInfo :: SourceInfo b <- asks getter
let customization = _siCustomization sourceInfo
tCase = _rscNamingConvention customization
-- insert in table
insertName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfInsert mkInsertField gqlName
-- insert one in table
insertOneName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfInsertOne mkInsertOneField gqlName
insert <- insertIntoTable backendInsertAction scenario tableInfo insertName insertDesc
Role-invariant schema constructors We build the GraphQL schema by combining building blocks such as `tableSelectionSet` and `columnParser`. These building blocks individually build `{InputFields,Field,}Parser` objects. Those object specify the valid GraphQL schema. Since the GraphQL schema is role-dependent, at some point we need to know what fragment of the GraphQL schema a specific role is allowed to access, and this is stored in `{Sel,Upd,Ins,Del}PermInfo` objects. We have passed around these permission objects as function arguments to the schema building blocks since we first started dealing with permissions during the PDV refactor - see hasura/graphql-engine@5168b99e463199b1934d8645bd6cd37eddb64ae1 in hasura/graphql-engine#4111. This means that, for instance, `tableSelectionSet` has as its type: ```haskell tableSelectionSet :: forall b r m n. MonadBuildSchema b r m n => SourceName -> TableInfo b -> SelPermInfo b -> m (Parser 'Output n (AnnotatedFields b)) ``` There are three reasons to change this. 1. We often pass a `Maybe (xPermInfo b)` instead of a proper `xPermInfo b`, and it's not clear what the intended semantics of this is. Some potential improvements on the data types involved are discussed in issue hasura/graphql-engine-mono#3125. 2. In most cases we also already pass a `TableInfo b`, and together with the `MonadRole` that is usually also in scope, this means that we could look up the required permissions regardless: so passing the permissions explicitly undermines the "single source of truth" principle. Breaking this principle also makes the code more difficult to read. 3. We are working towards role-based parsers (see hasura/graphql-engine-mono#2711), where the `{InputFields,Field,}Parser` objects are constructed in a role-invariant way, so that we have a single object that can be used for all roles. In particular, this means that the schema building blocks _need_ to be constructed in a role-invariant way. While this PR doesn't accomplish that, it does reduce the amount of role-specific arguments being passed, thus fixing hasura/graphql-engine-mono#3068. Concretely, this PR simply drops the `xPermInfo b` argument from almost all schema building blocks. Instead these objects are looked up from the `TableInfo b` as-needed. The resulting code is considerably simpler and shorter. One way to interpret this change is as follows. Before this PR, we figured out permissions at the top-level in `Hasura.GraphQL.Schema`, passing down the obtained `xPermInfo` objects as required. After this PR, we have a bottom-up approach where the schema building blocks themselves decide whether they want to be included for a particular role. So this moves some permission logic out of `Hasura.GraphQL.Schema`, which is very complex. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3608 GitOrigin-RevId: 51a744f34ec7d57bc8077667ae7f9cb9c4f6c962
2022-02-17 11:16:20 +03:00
-- Select permissions are required for insertOne: the selection set is the
-- same as a select on that table, and it therefore can't be populated if the
-- user doesn't have select permissions.
insertOne <- insertOneIntoTable backendInsertAction scenario tableInfo insertOneName insertOneDesc
Role-invariant schema constructors We build the GraphQL schema by combining building blocks such as `tableSelectionSet` and `columnParser`. These building blocks individually build `{InputFields,Field,}Parser` objects. Those object specify the valid GraphQL schema. Since the GraphQL schema is role-dependent, at some point we need to know what fragment of the GraphQL schema a specific role is allowed to access, and this is stored in `{Sel,Upd,Ins,Del}PermInfo` objects. We have passed around these permission objects as function arguments to the schema building blocks since we first started dealing with permissions during the PDV refactor - see hasura/graphql-engine@5168b99e463199b1934d8645bd6cd37eddb64ae1 in hasura/graphql-engine#4111. This means that, for instance, `tableSelectionSet` has as its type: ```haskell tableSelectionSet :: forall b r m n. MonadBuildSchema b r m n => SourceName -> TableInfo b -> SelPermInfo b -> m (Parser 'Output n (AnnotatedFields b)) ``` There are three reasons to change this. 1. We often pass a `Maybe (xPermInfo b)` instead of a proper `xPermInfo b`, and it's not clear what the intended semantics of this is. Some potential improvements on the data types involved are discussed in issue hasura/graphql-engine-mono#3125. 2. In most cases we also already pass a `TableInfo b`, and together with the `MonadRole` that is usually also in scope, this means that we could look up the required permissions regardless: so passing the permissions explicitly undermines the "single source of truth" principle. Breaking this principle also makes the code more difficult to read. 3. We are working towards role-based parsers (see hasura/graphql-engine-mono#2711), where the `{InputFields,Field,}Parser` objects are constructed in a role-invariant way, so that we have a single object that can be used for all roles. In particular, this means that the schema building blocks _need_ to be constructed in a role-invariant way. While this PR doesn't accomplish that, it does reduce the amount of role-specific arguments being passed, thus fixing hasura/graphql-engine-mono#3068. Concretely, this PR simply drops the `xPermInfo b` argument from almost all schema building blocks. Instead these objects are looked up from the `TableInfo b` as-needed. The resulting code is considerably simpler and shorter. One way to interpret this change is as follows. Before this PR, we figured out permissions at the top-level in `Hasura.GraphQL.Schema`, passing down the obtained `xPermInfo` objects as required. After this PR, we have a bottom-up approach where the schema building blocks themselves decide whether they want to be included for a particular role. So this moves some permission logic out of `Hasura.GraphQL.Schema`, which is very complex. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3608 GitOrigin-RevId: 51a744f34ec7d57bc8077667ae7f9cb9c4f6c962
2022-02-17 11:16:20 +03:00
pure $ catMaybes [insert, insertOne]
where
insertDesc = buildFieldDescription defaultInsertDesc $ _crfComment _tcrfInsert
insertOneDesc = buildFieldDescription defaultInsertOneDesc $ _crfComment _tcrfInsertOne
defaultInsertDesc = "insert data into the table: " <>> tableName
defaultInsertOneDesc = "insert a single row into the table: " <>> tableName
TableCustomRootFields {..} = _tcCustomRootFields . _tciCustomConfig $ _tiCoreInfo tableInfo
-- | This function is the basic building block for update mutations. It
-- implements the mutation schema in the general shape described in
-- @https://hasura.io/docs/latest/graphql/core/databases/postgres/mutations/update.html@.
--
-- Something that varies between backends is the @update operators@ that they
-- support (i.e. the schema fields @_set@, @_inc@, etc., see
-- <src/Hasura.Backends.Postgres.Instances.Schema.html#updateOperators Hasura.Backends.Postgres.Instances.Schema.updateOperators> for an example
-- implementation). Therefore, this function is parameterised over a monadic
-- action that produces the operators that the backend supports in the context
-- of some table and associated update permissions.
--
-- Apart from this detail, the rest of the arguments are the same as those
-- of @BackendSchema.@'Hasura.GraphQL.Schema.Backend.buildTableUpdateMutationFields'.
--
-- The suggested way to use this is like:
--
-- > instance BackendSchema MyBackend where
-- > ...
-- > buildTableUpdateMutationFields = GSB.buildTableUpdateMutationFields myBackendUpdateOperators
-- > ...
buildTableUpdateMutationFields ::
forall b r m n.
( MonadBuildSchema b r m n,
AggregationPredicatesSchema b,
BackendTableSelectSchema b
) =>
-- | an action that builds @BackendUpdate@ with the
-- backend-specific data needed to perform an update mutation
( TableInfo b ->
SchemaT
r
m
(InputFieldsParser n (BackendUpdate b (UnpreparedValue b)))
) ->
MkRootFieldName ->
Scenario ->
-- | The name of the table being acted on
TableName b ->
-- | table info
TableInfo b ->
-- | field display name
C.GQLNameIdentifier ->
SchemaT r m [FieldParser n (AnnotatedUpdateG b (RemoteRelationshipField UnpreparedValue) (UnpreparedValue b))]
buildTableUpdateMutationFields mkBackendUpdate mkRootFieldName scenario tableName tableInfo gqlName = do
sourceInfo :: SourceInfo b <- asks getter
let customization = _siCustomization sourceInfo
tCase = _rscNamingConvention customization
-- update table
updateName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfUpdate mkUpdateField gqlName
-- update table by pk
updatePKName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfUpdateByPk mkUpdateByPkField gqlName
backendUpdate <- mkBackendUpdate tableInfo
update <- updateTable backendUpdate scenario tableInfo updateName updateDesc
-- Primary keys can only be tested in the `where` clause if a primary key
-- exists on the table and if the user has select permissions on all columns
-- that make up the key.
updateByPk <- updateTableByPk backendUpdate scenario tableInfo updatePKName updatePKDesc
Role-invariant schema constructors We build the GraphQL schema by combining building blocks such as `tableSelectionSet` and `columnParser`. These building blocks individually build `{InputFields,Field,}Parser` objects. Those object specify the valid GraphQL schema. Since the GraphQL schema is role-dependent, at some point we need to know what fragment of the GraphQL schema a specific role is allowed to access, and this is stored in `{Sel,Upd,Ins,Del}PermInfo` objects. We have passed around these permission objects as function arguments to the schema building blocks since we first started dealing with permissions during the PDV refactor - see hasura/graphql-engine@5168b99e463199b1934d8645bd6cd37eddb64ae1 in hasura/graphql-engine#4111. This means that, for instance, `tableSelectionSet` has as its type: ```haskell tableSelectionSet :: forall b r m n. MonadBuildSchema b r m n => SourceName -> TableInfo b -> SelPermInfo b -> m (Parser 'Output n (AnnotatedFields b)) ``` There are three reasons to change this. 1. We often pass a `Maybe (xPermInfo b)` instead of a proper `xPermInfo b`, and it's not clear what the intended semantics of this is. Some potential improvements on the data types involved are discussed in issue hasura/graphql-engine-mono#3125. 2. In most cases we also already pass a `TableInfo b`, and together with the `MonadRole` that is usually also in scope, this means that we could look up the required permissions regardless: so passing the permissions explicitly undermines the "single source of truth" principle. Breaking this principle also makes the code more difficult to read. 3. We are working towards role-based parsers (see hasura/graphql-engine-mono#2711), where the `{InputFields,Field,}Parser` objects are constructed in a role-invariant way, so that we have a single object that can be used for all roles. In particular, this means that the schema building blocks _need_ to be constructed in a role-invariant way. While this PR doesn't accomplish that, it does reduce the amount of role-specific arguments being passed, thus fixing hasura/graphql-engine-mono#3068. Concretely, this PR simply drops the `xPermInfo b` argument from almost all schema building blocks. Instead these objects are looked up from the `TableInfo b` as-needed. The resulting code is considerably simpler and shorter. One way to interpret this change is as follows. Before this PR, we figured out permissions at the top-level in `Hasura.GraphQL.Schema`, passing down the obtained `xPermInfo` objects as required. After this PR, we have a bottom-up approach where the schema building blocks themselves decide whether they want to be included for a particular role. So this moves some permission logic out of `Hasura.GraphQL.Schema`, which is very complex. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3608 GitOrigin-RevId: 51a744f34ec7d57bc8077667ae7f9cb9c4f6c962
2022-02-17 11:16:20 +03:00
pure $ catMaybes [update, updateByPk]
where
updateDesc = buildFieldDescription defaultUpdateDesc $ _crfComment _tcrfUpdate
updatePKDesc = buildFieldDescription defaultUpdatePKDesc $ _crfComment _tcrfUpdateByPk
defaultUpdateDesc = "update data of the table: " <>> tableName
defaultUpdatePKDesc = "update single row of the table: " <>> tableName
TableCustomRootFields {..} = _tcCustomRootFields . _tciCustomConfig $ _tiCoreInfo tableInfo
buildTableDeleteMutationFields ::
forall b r m n.
( MonadBuildSchema b r m n,
AggregationPredicatesSchema b,
BackendTableSelectSchema b
) =>
MkRootFieldName ->
Scenario ->
TableName b ->
TableInfo b ->
C.GQLNameIdentifier ->
SchemaT r m [FieldParser n (AnnDelG b (RemoteRelationshipField UnpreparedValue) (UnpreparedValue b))]
buildTableDeleteMutationFields mkRootFieldName scenario tableName tableInfo gqlName = do
sourceInfo :: SourceInfo b <- asks getter
let customization = _siCustomization sourceInfo
tCase = _rscNamingConvention customization
-- delete from table
deleteName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfDelete mkDeleteField gqlName
-- delete from table by pk
deletePKName = runMkRootFieldName mkRootFieldName $ setFieldNameCase tCase tableInfo _tcrfDeleteByPk mkDeleteByPkField gqlName
delete <- deleteFromTable scenario tableInfo deleteName deleteDesc
-- Primary keys can only be tested in the `where` clause if the user has
-- select permissions for them, which at the very least requires select
-- permissions.
deleteByPk <- deleteFromTableByPk scenario tableInfo deletePKName deletePKDesc
Role-invariant schema constructors We build the GraphQL schema by combining building blocks such as `tableSelectionSet` and `columnParser`. These building blocks individually build `{InputFields,Field,}Parser` objects. Those object specify the valid GraphQL schema. Since the GraphQL schema is role-dependent, at some point we need to know what fragment of the GraphQL schema a specific role is allowed to access, and this is stored in `{Sel,Upd,Ins,Del}PermInfo` objects. We have passed around these permission objects as function arguments to the schema building blocks since we first started dealing with permissions during the PDV refactor - see hasura/graphql-engine@5168b99e463199b1934d8645bd6cd37eddb64ae1 in hasura/graphql-engine#4111. This means that, for instance, `tableSelectionSet` has as its type: ```haskell tableSelectionSet :: forall b r m n. MonadBuildSchema b r m n => SourceName -> TableInfo b -> SelPermInfo b -> m (Parser 'Output n (AnnotatedFields b)) ``` There are three reasons to change this. 1. We often pass a `Maybe (xPermInfo b)` instead of a proper `xPermInfo b`, and it's not clear what the intended semantics of this is. Some potential improvements on the data types involved are discussed in issue hasura/graphql-engine-mono#3125. 2. In most cases we also already pass a `TableInfo b`, and together with the `MonadRole` that is usually also in scope, this means that we could look up the required permissions regardless: so passing the permissions explicitly undermines the "single source of truth" principle. Breaking this principle also makes the code more difficult to read. 3. We are working towards role-based parsers (see hasura/graphql-engine-mono#2711), where the `{InputFields,Field,}Parser` objects are constructed in a role-invariant way, so that we have a single object that can be used for all roles. In particular, this means that the schema building blocks _need_ to be constructed in a role-invariant way. While this PR doesn't accomplish that, it does reduce the amount of role-specific arguments being passed, thus fixing hasura/graphql-engine-mono#3068. Concretely, this PR simply drops the `xPermInfo b` argument from almost all schema building blocks. Instead these objects are looked up from the `TableInfo b` as-needed. The resulting code is considerably simpler and shorter. One way to interpret this change is as follows. Before this PR, we figured out permissions at the top-level in `Hasura.GraphQL.Schema`, passing down the obtained `xPermInfo` objects as required. After this PR, we have a bottom-up approach where the schema building blocks themselves decide whether they want to be included for a particular role. So this moves some permission logic out of `Hasura.GraphQL.Schema`, which is very complex. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3608 GitOrigin-RevId: 51a744f34ec7d57bc8077667ae7f9cb9c4f6c962
2022-02-17 11:16:20 +03:00
pure $ catMaybes [delete, deleteByPk]
where
deleteDesc = buildFieldDescription defaultDeleteDesc $ _crfComment _tcrfDelete
deletePKDesc = buildFieldDescription defaultDeletePKDesc $ _crfComment _tcrfDeleteByPk
defaultDeleteDesc = "delete data from the table: " <>> tableName
defaultDeletePKDesc = "delete single row from the table: " <>> tableName
TableCustomRootFields {..} = _tcCustomRootFields . _tciCustomConfig $ _tiCoreInfo tableInfo
buildFieldDescription :: Text -> Comment -> Maybe G.Description
buildFieldDescription defaultDescription = \case
Automatic -> Just $ G.Description defaultDescription
Explicit comment -> G.Description . toTxt <$> comment