graphql-engine/server/src-lib/Hasura/RQL/DDL/Schema/Cache/Dependencies.hs
Karthikeyan Chinnakonda 92026b769f [Preview] Inherited roles for postgres read queries
fixes #3868

docker image - `hasura/graphql-engine:inherited-roles-preview-48b73a2de`

Note:

To be able to use the inherited roles feature, the graphql-engine should be started with the env variable `HASURA_GRAPHQL_EXPERIMENTAL_FEATURES` set to `inherited_roles`.

Introduction
------------

This PR implements the idea of multiple roles as presented in this [paper](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/FGALanguageICDE07.pdf). The multiple roles feature in this PR can be used via inherited roles. An inherited role is a role which can be created by combining multiple singular roles. For example, if there are two roles `author` and `editor` configured in the graphql-engine, then we can create a inherited role with the name of `combined_author_editor` role which will combine the select permissions of the `author` and `editor` roles and then make GraphQL queries using the `combined_author_editor`.

How are select permissions of different roles are combined?
------------------------------------------------------------

A select permission includes 5 things:

1. Columns accessible to the role
2. Row selection filter
3. Limit
4. Allow aggregation
5. Scalar computed fields accessible to the role

 Suppose there are two roles, `role1` gives access to the `address` column with row filter `P1` and `role2` gives access to both the `address` and the `phone` column with row filter `P2` and we create a new role `combined_roles` which combines `role1` and `role2`.

Let's say the following GraphQL query is queried with the `combined_roles` role.

```graphql
query {
   employees {
     address
     phone
   }
}
```

This will translate to the following SQL query:

```sql

 select
    (case when (P1 or P2) then address else null end) as address,
    (case when P2 then phone else null end) as phone
 from employee
 where (P1 or P2)
```

The other parameters of the select permission will be combined in the following manner:

1. Limit - Minimum of the limits will be the limit of the inherited role
2. Allow aggregations - If any of the role allows aggregation, then the inherited role will allow aggregation
3. Scalar computed fields - same as table column fields, as in the above example

APIs for inherited roles:
----------------------

1. `add_inherited_role`

`add_inherited_role` is the [metadata API](https://hasura.io/docs/1.0/graphql/core/api-reference/index.html#schema-metadata-api) to create a new inherited role. It accepts two arguments

`role_name`: the name of the inherited role to be added (String)
`role_set`: list of roles that need to be combined (Array of Strings)

Example:

```json
{
  "type": "add_inherited_role",
  "args": {
      "role_name":"combined_user",
      "role_set":[
          "user",
          "user1"
      ]
  }
}
```

After adding the inherited role, the inherited role can be used like single roles like earlier

Note:

An inherited role can only be created with non-inherited/singular roles.

2. `drop_inherited_role`

The `drop_inherited_role` API accepts the name of the inherited role and drops it from the metadata. It accepts a single argument:

`role_name`: name of the inherited role to be dropped

Example:

```json

{
  "type": "drop_inherited_role",
  "args": {
      "role_name":"combined_user"
  }
}
```

Metadata
---------

The derived roles metadata will be included under the `experimental_features` key while exporting the metadata.

```json
{
  "experimental_features": {
    "derived_roles": [
      {
        "role_name": "manager_is_employee_too",
        "role_set": [
          "employee",
          "manager"
        ]
      }
    ]
  }
}
```

Scope
------

Only postgres queries and subscriptions are supported in this PR.

Important points:
-----------------

1. All columns exposed to an inherited role will be marked as `nullable`, this is done so that cell value nullification can be done.

TODOs
-------

- [ ] Tests
   - [ ] Test a GraphQL query running with a inherited role without enabling inherited roles in experimental features
   - [] Tests for aggregate queries, limit, computed fields, functions, subscriptions (?)
   - [ ] Introspection test with a inherited role (nullability changes in a inherited role)
- [ ] Docs
- [ ] Changelog

Co-authored-by: Vamshi Surabhi <6562944+0x777@users.noreply.github.com>
GitOrigin-RevId: 3b8ee1e11f5ceca80fe294f8c074d42fbccfec63
2021-03-08 11:15:10 +00:00

188 lines
10 KiB
Haskell

module Hasura.RQL.DDL.Schema.Cache.Dependencies
( resolveDependencies
) where
import Hasura.Prelude
import qualified Data.HashMap.Strict.Extended as M
import qualified Data.HashSet as HS
import Control.Arrow.Extended
import Control.Lens hiding ((.=))
import Data.Aeson
import Data.List (nub)
import Data.Monoid (First)
import Data.Text.Extended
import Data.Typeable (cast)
import Hasura.RQL.DDL.Schema.Cache.Common
import Hasura.RQL.Types
-- | Processes collected 'CIDependency' values into a 'DepMap', performing integrity checking to
-- ensure the dependencies actually exist. If a dependency is missing, its transitive dependents are
-- removed from the cache, and 'InconsistentMetadata's are returned.
resolveDependencies
:: (ArrowKleisli m arr, QErrM m)
=> ( BuildOutputs
, [(MetadataObject, SchemaObjId, SchemaDependency)]
) `arr` (BuildOutputs, [InconsistentMetadata], DepMap)
resolveDependencies = arrM \(cache, dependencies) -> do
let dependencyMap = dependencies
& M.groupOn (view _2)
& fmap (map \(metadataObject, _, schemaDependency) -> (metadataObject, schemaDependency))
performIteration 0 cache [] dependencyMap
-- Processes dependencies using an iterative process that alternates between two steps:
--
-- 1. First, pruneDanglingDependents searches for any dependencies that do not exist in the
-- current cache and removes their dependents from the dependency map, returning an
-- InconsistentMetadata for each dependent that was removed. This step does not change
-- the schema cache in any way.
--
-- 2. Second, deleteMetadataObject drops the pruned dependent objects from the cache. It does
-- not alter (or consult) the dependency map, so transitive dependents are /not/ removed.
--
-- By iterating the above process until pruneDanglingDependents does not discover any new
-- inconsistencies, all missing dependencies will eventually be removed, and since dependency
-- graphs between schema objects are unlikely to be very deep, it will usually terminate in just
-- a few iterations.
performIteration
:: (QErrM m)
=> Int
-> BuildOutputs
-> [InconsistentMetadata]
-> HashMap SchemaObjId [(MetadataObject, SchemaDependency)]
-> m (BuildOutputs, [InconsistentMetadata], DepMap)
performIteration iterationNumber cache inconsistencies dependencies = do
let (newInconsistencies, prunedDependencies) = pruneDanglingDependents cache dependencies
case newInconsistencies of
[] -> pure (cache, inconsistencies, HS.fromList . map snd <$> prunedDependencies)
_ | iterationNumber < 100 -> do
let inconsistentIds = nub $ concatMap imObjectIds newInconsistencies
prunedCache = foldl' (flip deleteMetadataObject) cache inconsistentIds
allInconsistencies = inconsistencies <> newInconsistencies
performIteration (iterationNumber + 1) prunedCache allInconsistencies prunedDependencies
| otherwise ->
-- Running for 100 iterations without terminating is (hopefully) enormously unlikely
-- unless we did something very wrong, so halt the process and abort with some
-- debugging information.
throwError (err500 Unexpected "schema dependency resolution failed to terminate")
{ qeInternal = Just $ object
[ "inconsistent_objects" .= object
[ "old" .= inconsistencies
, "new" .= newInconsistencies ]
, "pruned_dependencies" .= (map snd <$> prunedDependencies) ] }
pruneDanglingDependents
:: BuildOutputs
-> HashMap SchemaObjId [(MetadataObject, SchemaDependency)]
-> ([InconsistentMetadata], HashMap SchemaObjId [(MetadataObject, SchemaDependency)])
pruneDanglingDependents cache = fmap (M.filter (not . null)) . traverse do
partitionEithers . map \(metadataObject, dependency) -> case resolveDependency dependency of
Right () -> Right (metadataObject, dependency)
Left errorMessage -> Left (InconsistentObject errorMessage metadataObject)
where
resolveDependency :: SchemaDependency -> Either Text ()
resolveDependency (SchemaDependency objectId _) = case objectId of
SOSource source -> void $ M.lookup source (_boSources cache)
`onNothing` Left ("no such source exists: " <>> source)
SORemoteSchema remoteSchemaName -> unless (remoteSchemaName `M.member` _boRemoteSchemas cache) $
Left $ "remote schema " <> remoteSchemaName <<> " is not found"
SORemoteSchemaPermission remoteSchemaName roleName -> do
remoteSchema <-
onNothing (M.lookup remoteSchemaName $ _boRemoteSchemas cache)
$ Left $ "remote schema " <> remoteSchemaName <<> " is not found"
unless (roleName `M.member` _rscPermissions (fst remoteSchema)) $
Left $ "no permission defined on remote schema " <> remoteSchemaName
<<> " for role " <>> roleName
SOSourceObj source sourceObjId -> do
sourceInfo <- castSourceInfo source sourceObjId
case sourceObjId of
SOITable tableName -> do
void $ resolveTable sourceInfo tableName
SOIFunction functionName -> void $
M.lookup functionName (_siFunctions sourceInfo)
`onNothing` Left ("function " <> functionName <<> " is not tracked")
SOITableObj tableName tableObjectId -> do
tableInfo <- resolveTable sourceInfo tableName
case tableObjectId of
TOCol columnName ->
void $ resolveField tableInfo (columnToFieldName tableInfo columnName) _FIColumn "column"
TORel relName ->
void $ resolveField tableInfo (fromRel relName) _FIRelationship "relationship"
TOComputedField fieldName ->
void $ resolveField tableInfo (fromComputedField fieldName) _FIComputedField "computed field"
TORemoteRel fieldName ->
void $ resolveField tableInfo (fromRemoteRelationship fieldName) _FIRemoteRelationship "remote relationship"
TOForeignKey constraintName -> do
let foreignKeys = _tciForeignKeys $ _tiCoreInfo tableInfo
unless (isJust $ find ((== constraintName) . _cName . _fkConstraint) foreignKeys) $
Left $ "no foreign key constraint named " <> constraintName <<> " is "
<> "defined for table " <>> tableName
TOPerm roleName permType -> withPermType permType \accessor -> do
let permLens = permAccToLens accessor
unless (has (tiRolePermInfoMap.ix roleName.permLens._Just) tableInfo) $
Left $ "no " <> permTypeToCode permType <> " permission defined on table "
<> tableName <<> " for role " <>> roleName
TOTrigger triggerName ->
unless (M.member triggerName (_tiEventTriggerInfoMap tableInfo)) $ Left $
"no event trigger named " <> triggerName <<> " is defined for table " <>> tableName
castSourceInfo
:: (Backend b) => SourceName -> SourceObjId b -> Either Text (SourceInfo b)
castSourceInfo sourceName _ =
-- TODO: if the cast returns Nothing, we should be throwing an internal error
-- the type of the dependency in sources is not as recorded
(M.lookup sourceName (_boSources cache) >>= \(BackendSourceInfo si) -> cast si)
`onNothing` Left ("no such source found " <>> sourceName)
resolveTable sourceInfo tableName =
M.lookup tableName (_siTables sourceInfo)
`onNothing` Left ("table " <> tableName <<> " is not tracked")
columnToFieldName :: forall b. (Backend b) => TableInfo b -> Column b -> FieldName
columnToFieldName _ = fromCol @b
resolveField
:: Backend b
=> TableInfo b -> FieldName -> Getting (First a) (FieldInfo b) a -> Text -> Either Text a
resolveField tableInfo fieldName fieldType fieldTypeName = do
let coreInfo = _tiCoreInfo tableInfo
tableName = _tciName coreInfo
fieldInfo <- M.lookup fieldName (_tciFieldInfoMap coreInfo) `onNothing` Left
("table " <> tableName <<> " has no field named " <>> fieldName)
(fieldInfo ^? fieldType) `onNothing` Left
("field " <> fieldName <<> "of table " <> tableName <<> " is not a " <> fieldTypeName)
deleteMetadataObject
:: MetadataObjId -> BuildOutputs -> BuildOutputs
deleteMetadataObject = \case
MOSource name -> boSources %~ M.delete name
MOSourceObjId source sourceObjId -> boSources %~ M.adjust (deleteObjId sourceObjId) source
MORemoteSchema name -> boRemoteSchemas %~ M.delete name
MORemoteSchemaPermissions name role -> boRemoteSchemas.ix name._1.rscPermissions %~ M.delete role
MOCronTrigger name -> boCronTriggers %~ M.delete name
MOCustomTypes -> boCustomTypes %~ const emptyAnnotatedCustomTypes
MOAction name -> boActions %~ M.delete name
MOEndpoint name -> boEndpoints %~ M.delete name
MOActionPermission name role -> boActions.ix name.aiPermissions %~ M.delete role
MOInheritedRole name -> boInheritedRoles %~ M.delete name
where
deleteObjId :: (Backend b) => SourceMetadataObjId b -> BackendSourceInfo -> BackendSourceInfo
deleteObjId sourceObjId sourceInfo = maybe sourceInfo (BackendSourceInfo . deleteObjFn sourceObjId) $ unsafeSourceInfo sourceInfo
deleteObjFn :: (Backend b) => SourceMetadataObjId b -> SourceInfo b -> SourceInfo b
deleteObjFn = \case
SMOTable name -> siTables %~ M.delete name
SMOFunction name -> siFunctions %~ M.delete name
SMOFunctionPermission functionName role ->
siFunctions.ix functionName.fiPermissions %~ HS.delete role
SMOTableObj tableName tableObjectId -> siTables.ix tableName %~ case tableObjectId of
MTORel name _ -> tiCoreInfo.tciFieldInfoMap %~ M.delete (fromRel name)
MTOComputedField name -> tiCoreInfo.tciFieldInfoMap %~ M.delete (fromComputedField name)
MTORemoteRelationship name -> tiCoreInfo.tciFieldInfoMap %~ M.delete (fromRemoteRelationship name)
MTOTrigger name -> tiEventTriggerInfoMap %~ M.delete name
MTOPerm roleName permType -> withPermType permType \accessor ->
tiRolePermInfoMap.ix roleName.permAccToLens accessor .~ Nothing