[#2401] Configs in schema files

CHANGELOG.md

@@ -1,8 +1,11 @@
 ## dbt 0.21.0 (Release TBD)
 
-## dbt 0.21.0b1 (August 03, 2021)
+### Features
+- Enable setting configs in schema files for models, seeds, snapshots, analyses, tests ([#2401](https://github.com/dbt-labs/dbt/issues/2401), [#3616](https://github.com/dbt-labs/dbt/pull/3616))
 
 
+## dbt 0.21.0b1 (August 03, 2021)
+
 ### Breaking changes
 - Add full node selection to source freshness command and align selection syntax with other tasks (`dbt source freshness --select source_name` --> `dbt source freshness --select source:souce_name`) and rename `dbt source snapshot-freshness` -> `dbt source freshness`. ([#2987](https://github.com/dbt-labs/dbt/issues/2987), [#3554](https://github.com/dbt-labs/dbt/pull/3554))
 - **dbt-snowflake:** Turn off transactions and turn on `autocommit` by default. Explicitly specify `begin` and `commit` for DML statements in incremental and snapshot materializations. Note that this may affect user-space code that depends on transactions.

core/dbt/context/context_config.py

@@ -125,6 +125,7 @@ def calculate_node_config(
         resource_type: NodeType,
         project_name: str,
         base: bool,
+        patch_config_dict: Dict[str, Any] = None
     ) -> BaseConfig:
         own_config = self.get_node_project(project_name)
 
@@ -134,6 +135,12 @@ def calculate_node_config(
         for fqn_config in project_configs:
             result = self._update_from_config(result, fqn_config)
 
+        # When schema files patch config, it has lower precedence than
+        # config in the models (config_call_dict), so we add the patch_config_dict
+        # before the config_call_dict
+        if patch_config_dict:
+            result = self._update_from_config(result, patch_config_dict)
+
         # config_calls are created in the 'experimental' model parser and
         # the ParseConfigObject (via add_config_call)
         result = self._update_from_config(result, config_call_dict)
@@ -153,6 +160,7 @@ def calculate_node_config_dict(
         resource_type: NodeType,
         project_name: str,
         base: bool,
+        patch_config_dict: Dict[str, Any],
     ) -> Dict[str, Any]:
         ...
 
@@ -192,13 +200,15 @@ def calculate_node_config_dict(
         resource_type: NodeType,
         project_name: str,
         base: bool,
+        patch_config_dict: dict = None
     ) -> Dict[str, Any]:
         config = self.calculate_node_config(
             config_call_dict=config_call_dict,
             fqn=fqn,
             resource_type=resource_type,
             project_name=project_name,
             base=base,
+            patch_config_dict=patch_config_dict
         )
         finalized = config.finalize_and_validate()
         return finalized.to_dict(omit_none=True)
@@ -215,13 +225,15 @@ def calculate_node_config_dict(
         resource_type: NodeType,
         project_name: str,
         base: bool,
+        patch_config_dict: dict = None
     ) -> Dict[str, Any]:
         return self.calculate_node_config(
             config_call_dict=config_call_dict,
             fqn=fqn,
             resource_type=resource_type,
             project_name=project_name,
             base=base,
+            patch_config_dict=patch_config_dict
         )
 
     def initial_result(
@@ -284,6 +296,7 @@ def build_config_dict(
         base: bool = False,
         *,
         rendered: bool = True,
+        patch_config_dict: dict = None
     ) -> Dict[str, Any]:
         if rendered:
             src = ContextConfigGenerator(self._active_project)
@@ -296,4 +309,5 @@ def build_config_dict(
             resource_type=self._resource_type,
             project_name=self._project_name,
             base=base,
+            patch_config_dict=patch_config_dict
         )

core/dbt/context/providers.py

@@ -1243,15 +1243,15 @@ class ModelContext(ProviderContext):
 
     @contextproperty
     def pre_hooks(self) -> List[Dict[str, Any]]:
-        if isinstance(self.model, ParsedSourceDefinition):
+        if self.model.resource_type in [NodeType.Source, NodeType.Test]:
             return []
         return [
             h.to_dict(omit_none=True) for h in self.model.config.pre_hook
         ]
 
     @contextproperty
     def post_hooks(self) -> List[Dict[str, Any]]:
-        if isinstance(self.model, ParsedSourceDefinition):
+        if self.model.resource_type in [NodeType.Source, NodeType.Test]:
             return []
         return [
             h.to_dict(omit_none=True) for h in self.model.config.post_hook

core/dbt/contracts/graph/compiled.py

@@ -109,15 +109,19 @@ class CompiledSnapshotNode(CompiledNode):
 @dataclass
 class CompiledDataTestNode(CompiledNode):
     resource_type: NodeType = field(metadata={'restrict': [NodeType.Test]})
-    config: TestConfig = field(default_factory=TestConfig)
+    # Was not able to make mypy happy and keep the code working. We need to
+    # refactor the various configs.
+    config: TestConfig = field(default_factory=TestConfig)  # type:ignore
 
 
 @dataclass
 class CompiledSchemaTestNode(CompiledNode, HasTestMetadata):
     # keep this in sync with ParsedSchemaTestNode!
     resource_type: NodeType = field(metadata={'restrict': [NodeType.Test]})
     column_name: Optional[str] = None
-    config: TestConfig = field(default_factory=TestConfig)
+    # Was not able to make mypy happy and keep the code working. We need to
+    # refactor the various configs.
+    config: TestConfig = field(default_factory=TestConfig)  # type:ignore
 
     def same_contents(self, other) -> bool:
         if other is None:

core/dbt/contracts/graph/manifest.py

@@ -14,7 +14,7 @@
     CompileResultNode, ManifestNode, NonSourceCompiledNode, GraphMemberNode
 )
 from dbt.contracts.graph.parsed import (
-    ParsedMacro, ParsedDocumentation, ParsedNodePatch, ParsedMacroPatch,
+    ParsedMacro, ParsedDocumentation,
     ParsedSourceDefinition, ParsedExposure, HasUniqueID,
     UnpatchedSourceDefinition, ManifestNodes
 )
@@ -26,9 +26,7 @@
 from dbt.dataclass_schema import dbtClassMixin
 from dbt.exceptions import (
     CompilationException,
-    raise_duplicate_resource_name, raise_compiler_error, warn_or_error,
-    raise_duplicate_patch_name,
-    raise_duplicate_macro_patch_name, raise_duplicate_source_patch_name,
+    raise_duplicate_resource_name, raise_compiler_error,
 )
 from dbt.helper_types import PathSet
 from dbt.logger import GLOBAL_LOGGER as logger
@@ -718,60 +716,6 @@ def get_resource_fqns(self) -> Mapping[str, PathSet]:
             resource_fqns[resource_type_plural].add(tuple(resource.fqn))
         return resource_fqns
 
-    # This is called by 'parse_patch' in the NodePatchParser
-    def add_patch(
-        self, source_file: SchemaSourceFile, patch: ParsedNodePatch,
-    ) -> None:
-        if patch.yaml_key in ['models', 'seeds', 'snapshots']:
-            unique_id = self.ref_lookup.get_unique_id(patch.name, None)
-        elif patch.yaml_key == 'analyses':
-            unique_id = self.analysis_lookup.get_unique_id(patch.name, None)
-        else:
-            raise dbt.exceptions.InternalException(
-                f'Unexpected yaml_key {patch.yaml_key} for patch in '
-                f'file {source_file.path.original_file_path}'
-            )
-        if unique_id is None:
-            # This will usually happen when a node is disabled
-            return
-
-        # patches can't be overwritten
-        node = self.nodes.get(unique_id)
-        if node:
-            if node.patch_path:
-                package_name, existing_file_path = node.patch_path.split('://')
-                raise_duplicate_patch_name(patch, existing_file_path)
-            source_file.append_patch(patch.yaml_key, unique_id)
-            node.patch(patch)
-
-    def add_macro_patch(
-        self, source_file: SchemaSourceFile, patch: ParsedMacroPatch,
-    ) -> None:
-        # macros are fully namespaced
-        unique_id = f'macro.{patch.package_name}.{patch.name}'
-        macro = self.macros.get(unique_id)
-        if not macro:
-            warn_or_error(
-                f'WARNING: Found documentation for macro "{patch.name}" '
-                f'which was not found'
-            )
-            return
-        if macro.patch_path:
-            package_name, existing_file_path = macro.patch_path.split('://')
-            raise_duplicate_macro_patch_name(patch, existing_file_path)
-        source_file.macro_patches[patch.name] = unique_id
-        macro.patch(patch)
-
-    def add_source_patch(
-        self, source_file: SchemaSourceFile, patch: SourcePatch,
-    ) -> None:
-        # source patches must be unique
-        key = (patch.overrides, patch.name)
-        if key in self.source_patches:
-            raise_duplicate_source_patch_name(patch, self.source_patches[key])
-        self.source_patches[key] = patch
-        source_file.source_patches.append(key)
-
     def get_used_schemas(self, resource_types=None):
         return frozenset({
             (node.database, node.schema) for node in

core/dbt/contracts/graph/model_config.py

@@ -268,9 +268,10 @@ def same_contents(
         return True
 
     # This is used in 'add_config_call' to created the combined config_call_dict.
+    # 'meta' moved here from node
     mergebehavior = {
         "append": ['pre-hook', 'pre_hook', 'post-hook', 'post_hook', 'tags'],
-        "update": ['quoting', 'column_types'],
+        "update": ['quoting', 'column_types', 'meta'],
     }
 
     @classmethod
@@ -355,10 +356,38 @@ class SourceConfig(BaseConfig):
 
 
 @dataclass
-class NodeConfig(BaseConfig):
+class NodeAndTestConfig(BaseConfig):
+    enabled: bool = True
+    # these fields are included in serialized output, but are not part of
+    # config comparison (they are part of database_representation)
+    alias: Optional[str] = field(
+        default=None,
+        metadata=CompareBehavior.Exclude.meta(),
+    )
+    schema: Optional[str] = field(
+        default=None,
+        metadata=CompareBehavior.Exclude.meta(),
+    )
+    database: Optional[str] = field(
+        default=None,
+        metadata=CompareBehavior.Exclude.meta(),
+    )
+    tags: Union[List[str], str] = field(
+        default_factory=list_str,
+        metadata=metas(ShowBehavior.Hide,
+                       MergeBehavior.Append,
+                       CompareBehavior.Exclude),
+    )
+    meta: Dict[str, Any] = field(
+        default_factory=dict,
+        metadata=MergeBehavior.Update.meta(),
+    )
+
+
+@dataclass
+class NodeConfig(NodeAndTestConfig):
     # Note: if any new fields are added with MergeBehavior, also update the
     # 'mergebehavior' dictionary
-    enabled: bool = True
     materialized: str = 'view'
     persist_docs: Dict[str, Any] = field(default_factory=dict)
     post_hook: List[Hook] = field(
@@ -369,11 +398,6 @@ class NodeConfig(BaseConfig):
         default_factory=list,
         metadata=MergeBehavior.Append.meta(),
     )
-    # this only applies for config v1, so it doesn't participate in comparison
-    vars: Dict[str, Any] = field(
-        default_factory=dict,
-        metadata=metas(CompareBehavior.Exclude, MergeBehavior.Update),
-    )
     quoting: Dict[str, Any] = field(
         default_factory=dict,
         metadata=MergeBehavior.Update.meta(),
@@ -384,26 +408,6 @@ class NodeConfig(BaseConfig):
         default_factory=dict,
         metadata=MergeBehavior.Update.meta(),
     )
-    # these fields are included in serialized output, but are not part of
-    # config comparison (they are part of database_representation)
-    alias: Optional[str] = field(
-        default=None,
-        metadata=CompareBehavior.Exclude.meta(),
-    )
-    schema: Optional[str] = field(
-        default=None,
-        metadata=CompareBehavior.Exclude.meta(),
-    )
-    database: Optional[str] = field(
-        default=None,
-        metadata=CompareBehavior.Exclude.meta(),
-    )
-    tags: Union[List[str], str] = field(
-        default_factory=list_str,
-        metadata=metas(ShowBehavior.Hide,
-                       MergeBehavior.Append,
-                       CompareBehavior.Exclude),
-    )
     full_refresh: Optional[bool] = None
     on_schema_change: Optional[str] = 'ignore'
 
@@ -447,7 +451,8 @@ class SeedConfig(NodeConfig):
 
 
 @dataclass
-class TestConfig(NodeConfig):
+class TestConfig(NodeAndTestConfig):
+    # this is repeated because of a different default
     schema: Optional[str] = field(
         default='dbt_test__audit',
         metadata=CompareBehavior.Exclude.meta(),