graphql-engine/server/src-lib/Hasura/GraphQL/Execute/Remote.hs

230 lines
8.4 KiB
Haskell
Raw Normal View History

{-# LANGUAGE DeriveAnyClass #-}
module Hasura.GraphQL.Execute.Remote
( buildExecStepRemote
, collectVariablesFromSelectionSet
, collectVariables
, resolveRemoteVariable
, resolveRemoteField
, runVariableCache
) where
import Hasura.Prelude
import qualified Data.Aeson as J
import qualified Data.HashMap.Strict as Map
import qualified Data.HashSet as Set
import qualified Data.Text as T
import qualified Language.GraphQL.Draft.Syntax as G
import Data.Text.Extended
import qualified Hasura.GraphQL.Transport.HTTP.Protocol as GH
import Hasura.Base.Error
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.GraphQL.Execute.Backend
import Hasura.GraphQL.Parser
import Hasura.GraphQL.Transport.HTTP.Protocol
import Hasura.RQL.Types
import Hasura.Session
mkVariableDefinitionAndValue :: Variable -> (G.VariableDefinition, (G.Name, J.Value))
mkVariableDefinitionAndValue var@(Variable varInfo gType varValue) =
(varDefn, (varName, varJSONValue))
where
varName = getName var
varDefn = G.VariableDefinition varName gType defaultVal
defaultVal =
case varInfo of
VIRequired _ -> Nothing
VIOptional _ val -> Just val
varJSONValue =
case varValue of
JSONValue v -> v
GraphQLValue val -> graphQLValueToJSON val
unresolveVariables
:: forall fragments
. Functor fragments
=> G.SelectionSet fragments Variable
-> G.SelectionSet fragments G.Name
unresolveVariables =
fmap (fmap (getName . vInfo))
collectVariables
:: forall fragments var
. (Foldable fragments, Hashable var, Eq var)
=> G.SelectionSet fragments var
-> Set.HashSet var
collectVariables =
Set.unions . fmap (foldMap Set.singleton)
collectVariablesFromSelectionSet
:: G.SelectionSet G.NoFragments Variable
-> [(G.VariableDefinition, (G.Name, J.Value))]
collectVariablesFromSelectionSet =
map mkVariableDefinitionAndValue . Set.toList . collectVariables
buildExecStepRemote
:: RemoteSchemaInfo
-> RemoteResultCustomizer
-> G.OperationType
-> G.SelectionSet G.NoFragments Variable
-> ExecutionStep
buildExecStepRemote remoteSchemaInfo resultCustomizer tp selSet =
let unresolvedSelSet = unresolveVariables selSet
allVars = map mkVariableDefinitionAndValue $ Set.toList $ collectVariables selSet
varValues = Map.fromList $ map snd allVars
varValsM = bool (Just varValues) Nothing $ Map.null varValues
varDefs = map fst allVars
_grQuery = G.TypedOperationDefinition tp Nothing varDefs [] unresolvedSelSet
_grVariables = varValsM
in ExecStepRemote remoteSchemaInfo resultCustomizer GH.GQLReq{_grOperationName = Nothing, ..}
-- | Association between keys uniquely identifying some remote JSON variable and
-- an 'Int' identifier that will be used to construct a valid variable name to
-- be used in a GraphQL query.
newtype RemoteJSONVariableMap =
RemoteJSONVariableMap (HashMap RemoteJSONVariableKey Int)
deriving newtype (Eq, Monoid, Semigroup)
-- | A unique identifier for some remote JSON variable whose name will need to
-- be substituted when constructing a GraphQL query.
--
-- For a detailed explanation of this behavior, see the following comment:
-- https://github.com/hasura/graphql-engine/issues/7170#issuecomment-880838970
data RemoteJSONVariableKey = RemoteJSONVariableKey !G.GType !J.Value
deriving stock (Eq, Generic)
deriving anyclass (Hashable)
-- | Resolves a `RemoteSchemaVariable` into a GraphQL `Variable`.
--
-- A `RemoteSchemaVariable` can either be a query variable (i.e. a variable
-- provided in the query) or it can be a `SessionPresetVariable` (in which case
-- we look up the value of the session variable and coerce it into the
-- appropriate type and then construct the GraphQL 'Variable').
--
-- NOTE: The session variable preset is a hard preset (i.e. if the session
-- variable doesn't exist, an error will be thrown).
--
-- The name of the GraphQL variable generated will be a GraphQL-ized version of
-- the session variable (i.e. '-' will be replaced with '_'), since session
-- variables are not valid GraphQL names.
--
-- Additionally, we need to handle partially traversed JSON values; likewise, we
-- create a new variable out of thin air.
--
-- For example, considering the following schema for a role:
--
-- input UserName {
-- firstName : String! @preset(value:"Foo")
-- lastName : String!
-- }
--
-- type Query {
-- user(
-- user_id: Int! @preset(value:"x-hasura-user-id")
-- user_name: UserName!
-- ): User
-- }
--
-- and the incoming query to the graphql-engine is:
--
-- query($foo: UserName!) {
-- user(user_name: $foo) { id name }
-- }
--
-- with variables:
--
-- { "foo": {"lastName": "Bar"} }
--
--
-- After resolving the session argument presets, the query that will be sent to
-- the remote server will be:
--
-- query ($x_hasura_user_id: Int!, $hasura_json_var_1: String!) {
-- user (user_id: $x_hasura_user_id, user_name: {firstName: "Foo", lastName: $hasura_json_var_1}) {
-- id
-- name
-- }
-- }
--
resolveRemoteVariable
:: (MonadError QErr m)
=> UserInfo
-> RemoteSchemaVariable
-> StateT RemoteJSONVariableMap m Variable
resolveRemoteVariable userInfo = \case
SessionPresetVariable sessionVar typeName presetInfo -> do
sessionVarVal <- onNothing (getSessionVariableValue sessionVar $ _uiSession userInfo)
$ throw400 NotFound $ sessionVar <<> " session variable expected, but not found"
let varName = sessionVariableToGraphQLName sessionVar
coercedValue <-
case presetInfo of
SessionArgumentPresetScalar ->
case G.unName typeName of
"Int" ->
case readMaybe $ T.unpack sessionVarVal of
Nothing -> throw400 CoercionError $ sessionVarVal <<> " cannot be coerced into an Int value"
Just i -> pure $ G.VInt i
"Boolean" ->
if | sessionVarVal `elem` ["true", "false"] ->
pure $ G.VBoolean $ "true" == sessionVarVal
| otherwise ->
throw400 CoercionError $ sessionVarVal <<> " cannot be coerced into a Boolean value"
"Float" ->
case readMaybe $ T.unpack sessionVarVal of
Nothing ->
throw400 CoercionError $ sessionVarVal <<> " cannot be coerced into a Float value"
Just i -> pure $ G.VFloat i
-- The `String`,`ID` and the default case all use the same code. But,
-- it will be better to not merge all of them into the default case
-- because it will be helpful to know how all the built-in scalars
-- are handled
"String" -> pure $ G.VString sessionVarVal
"ID" -> pure $ G.VString sessionVarVal
-- When we encounter a custom scalar, we just pass it as a string
_ -> pure $ G.VString sessionVarVal
SessionArgumentPresetEnum enumVals -> do
sessionVarEnumVal <-
G.EnumValue <$>
onNothing
(G.mkName sessionVarVal)
(throw400 CoercionError $ sessionVarVal <<> " is not a valid GraphQL name")
case sessionVarEnumVal `Set.member` enumVals of
True -> pure $ G.VEnum sessionVarEnumVal
False -> throw400 CoercionError $ sessionVarEnumVal <<> " is not one of the valid enum values"
-- nullability is false, because we treat presets as hard presets
let variableGType = G.TypeNamed (G.Nullability False) typeName
pure $ Variable (VIRequired varName) variableGType (GraphQLValue coercedValue)
RemoteJSONValue gtype jsonValue -> do
let key = RemoteJSONVariableKey gtype jsonValue
varMap <- gets coerce
index <- Map.lookup key varMap `onNothing` do
let i = Map.size varMap + 1
put . coerce $ Map.insert key i varMap
pure i
let varName = G.unsafeMkName $ "hasura_json_var_" <> tshow index
pure $ Variable (VIRequired varName) gtype $ JSONValue jsonValue
QueryVariable variable -> pure variable
-- | TODO: Documentation.
resolveRemoteField
:: (MonadError QErr m, Traversable f)
=> UserInfo
-> RemoteFieldG f RemoteSchemaVariable
-> StateT RemoteJSONVariableMap m (RemoteFieldG f Variable)
resolveRemoteField userInfo = traverse (resolveRemoteVariable userInfo)
-- | TODO: Documentation.
runVariableCache
:: Monad m
=> StateT RemoteJSONVariableMap m a
-> m a
runVariableCache = flip evalStateT mempty