graphql-engine/server/src-lib/Hasura/Backends/Postgres/Schema/OnConflict.hs
Auke Booij f026d44438 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@5168b99e46 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 08:17:17 +00:00

126 lines
5.6 KiB
Haskell

{-# LANGUAGE ApplicativeDo #-}
-- | Postgres Schema OnConflict
--
-- This module contains the building blocks for parsing @on_conflict@ clauses,
-- which in the Postgres backend are used to implement upsert functionality.
-- These are used by 'Hasura.Backends.Postgres.Instances.Schema.backendInsertParser' to
-- construct a postgres-specific schema parser for insert (and upsert) mutations.
module Hasura.Backends.Postgres.Schema.OnConflict
( onConflictFieldParser,
)
where
import Data.Text.Extended
import Hasura.GraphQL.Parser
( InputFieldsParser,
Kind (..),
Parser,
UnpreparedValue (..),
)
import Hasura.GraphQL.Parser qualified as P
import Hasura.GraphQL.Parser.Class
import Hasura.GraphQL.Schema.Backend
import Hasura.GraphQL.Schema.BoolExp
import Hasura.GraphQL.Schema.Common
import Hasura.GraphQL.Schema.Table
import Hasura.Prelude
import Hasura.RQL.IR.Insert qualified as IR
import Hasura.RQL.Types
import Language.GraphQL.Draft.Syntax qualified as G
-- | Parser for a field name @on_conflict@ of type @tablename_on_conflict@.
--
-- The @tablename_on_conflict@ object is used to generate the @ON CONFLICT@
-- SQL clause, indicating what should be done if an insert raises a conflict.
--
-- The types ordinarily produced by this parser are only created if the table has
-- unique or primary keys constraints.
--
-- If there are no columns for which the current role has update permissions, we
-- must still accept an empty list for @update_columns@ in the name of
-- backwards compatibility. We do this by adding a placeholder value to the
-- enum. See <https://github.com/hasura/graphql-engine/issues/6804>.
onConflictFieldParser ::
forall pgKind r m n.
MonadBuildSchema ('Postgres pgKind) r m n =>
SourceName ->
TableInfo ('Postgres pgKind) ->
m (InputFieldsParser n (Maybe (IR.OnConflictClause ('Postgres pgKind) (UnpreparedValue ('Postgres pgKind)))))
onConflictFieldParser sourceName tableInfo = do
updatePerms <- (_permUpd =<<) <$> tablePermissions tableInfo
let maybeConstraints = tciUniqueOrPrimaryKeyConstraints . _tiCoreInfo $ tableInfo
let maybeConflictObject = conflictObjectParser sourceName tableInfo <$> maybeConstraints <*> updatePerms
case maybeConflictObject of
Just conflictObject -> conflictObject <&> P.fieldOptional $$(G.litName "on_conflict") (Just "upsert condition")
Nothing -> return $ pure Nothing
-- | Create a parser for the @_on_conflict@ object of the given table.
conflictObjectParser ::
forall pgKind r m n.
MonadBuildSchema ('Postgres pgKind) r m n =>
SourceName ->
TableInfo ('Postgres pgKind) ->
NonEmpty (Constraint ('Postgres pgKind)) ->
UpdPermInfo ('Postgres pgKind) ->
m (Parser 'Input n (IR.OnConflictClause ('Postgres pgKind) (UnpreparedValue ('Postgres pgKind))))
conflictObjectParser sourceName tableInfo constraints updatePerms = do
updateColumnsEnum <- updateColumnsPlaceholderParser tableInfo
constraintParser <- conflictConstraint constraints sourceName tableInfo
whereExpParser <- boolExp sourceName tableInfo
tableGQLName <- getTableGQLName tableInfo
objectName <- P.mkTypename $ tableGQLName <> $$(G.litName "_on_conflict")
let presetColumns = partialSQLExpToUnpreparedValue <$> upiSet updatePerms
updateFilter = fmap partialSQLExpToUnpreparedValue <$> upiFilter updatePerms
objectDesc = G.Description $ "on_conflict condition type for table " <>> tableName
constraintName = $$(G.litName "constraint")
columnsName = $$(G.litName "update_columns")
whereExpName = $$(G.litName "where")
pure $
P.object objectName (Just objectDesc) $ do
constraint <- IR.CTConstraint <$> P.field constraintName Nothing constraintParser
whereExp <- P.fieldOptional whereExpName Nothing whereExpParser
updateColumns <-
P.fieldWithDefault columnsName Nothing (G.VList []) (P.list updateColumnsEnum) `P.bindFields` \cs ->
-- this can only happen if the placeholder was used
sequenceA cs `onNothing` parseError "erroneous column name"
pure $
case updateColumns of
[] -> IR.OCCDoNothing $ Just constraint
_ -> IR.OCCUpdate $ IR.OnConflictClauseData constraint updateColumns presetColumns $ BoolAnd $ updateFilter : maybeToList whereExp
where
tableName = tableInfoName tableInfo
-- | Constructs a Parser for the name of the constraints on a given table.
--
-- The TableCoreInfo of a given table contains a list of unique or primary key
-- constraints. Given the list of such constraints, this function creates a
-- parser for an enum type that matches it. This function makes no attempt at
-- de-duplicating contraint names, and assumes they are correct.
--
-- This function can fail if a constraint has a name that cannot be translated
-- to a GraphQL name (see hasura/graphql-engine-mono#1748).
conflictConstraint ::
forall pgKind r m n.
MonadBuildSchema ('Postgres pgKind) r m n =>
NonEmpty (Constraint ('Postgres pgKind)) ->
SourceName ->
TableInfo ('Postgres pgKind) ->
m (Parser 'Both n (ConstraintName ('Postgres pgKind)))
conflictConstraint constraints sourceName tableInfo =
memoizeOn 'conflictConstraint (sourceName, tableName) $ do
tableGQLName <- getTableGQLName tableInfo
constraintEnumValues <- for constraints \constraint -> do
name <- textToName $ toTxt $ _cName constraint
pure
( P.Definition name (Just "unique or primary key constraint") P.EnumValueInfo,
_cName constraint
)
enumName <- P.mkTypename $ tableGQLName <> $$(G.litName "_constraint")
let enumDesc = G.Description $ "unique or primary key constraints on table " <>> tableName
pure $ P.enum enumName (Just enumDesc) constraintEnumValues
where
tableName = tableInfoName tableInfo