airbyte-source-google-search-console 1.7.0__py3-none-any.whl → 1.9.0__py3-none-any.whl

{airbyte_source_google_search_console-1.7.0.dist-info → airbyte_source_google_search_console-1.9.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: airbyte-source-google-search-console
-Version: 1.7.0
+Version: 1.9.0
 Summary: Source implementation for Google Search Console.
 License: Elv2
 Author: Airbyte

airbyte_source_google_search_console-1.9.0.dist-info/RECORD

@@ -0,0 +1,13 @@
+source_google_search_console/__init__.py,sha256=HQCPu-CK7XmVDtP9rmTdB2XyraVCc6pv9pw38-O8y48,1191
+source_google_search_console/components.py,sha256=_6gl-xgwmvRgChEGNZhMsuYAcTw7gIf3yfe7IKvWkPk,6767
+source_google_search_console/config_migrations.py,sha256=Cl4SUdJpAf6wMM_vVhqjjU89NfUq9LIGJ9zNrWiBk-A,4235
+source_google_search_console/exceptions.py,sha256=iD3jYC4WxVCEKGsqQ7Vaj1tbjhJZ4S5mnSDnwFJdsIQ,1097
+source_google_search_console/manifest.yaml,sha256=s7WVxmLhgyobNNovFcHAu6JYKX_4UlqnSID1pC1HKJE,54144
+source_google_search_console/run.py,sha256=TBkPlseTERarkj6wL8AMEKgm5Xsb2drnltPVH6257-M,2195
+source_google_search_console/service_account_authenticator.py,sha256=pAWKAXfwfTY3xkXvQJH0EyFphFULdCIcC47YXYTO9X8,1307
+source_google_search_console/source.py,sha256=TFuafbBg8Nlb-LsYwNSXH1Xcz9U0ELZRX0e3hjCKPzs,7669
+source_google_search_console/spec.json,sha256=HebTMHop7twcJH6jjgAccYEgg93bTLGsp6jVdMj1x0c,9409
+airbyte_source_google_search_console-1.9.0.dist-info/METADATA,sha256=dqntQEZ33A7CkZ8YeSjbWtpglyVDRcw_4328p8Tkxjg,5621
+airbyte_source_google_search_console-1.9.0.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+airbyte_source_google_search_console-1.9.0.dist-info/entry_points.txt,sha256=DMcgc9bCX-Vt6hm_68pa77qS3eGdeMhg-UdlFc-XKUM,85
+airbyte_source_google_search_console-1.9.0.dist-info/RECORD,,

source_google_search_console/components.py

@@ -2,10 +2,13 @@
 # Copyright (c) 2025 Airbyte, Inc., all rights reserved.
 #
 
-from dataclasses import dataclass
-from typing import Any, Mapping, Optional
+from dataclasses import dataclass, field
+from typing import Any, Dict, List, Mapping, Optional
 
 from airbyte_cdk.sources.declarative.migrations.state_migration import StateMigration
+from airbyte_cdk.sources.declarative.schema import SchemaLoader
+from airbyte_cdk.sources.declarative.transformations import RecordTransformation
+from airbyte_cdk.sources.types import Config, StreamSlice, StreamState
 
 
 @dataclass
@@ -59,7 +62,6 @@ class NestedSubstreamStateMigration(StateMigration):
         return len(stream_state) > 0 and "states" not in stream_state
 
     def migrate(self, stream_state: Mapping[str, Any]) -> Mapping[str, Any]:
-        global_state: Optional[Mapping[str, Any]] = None
         per_partition_state = []
         for site_url_key, search_type_state in stream_state.items():
             if site_url_key == "date":
@@ -79,3 +81,103 @@ class NestedSubstreamStateMigration(StateMigration):
             "use_global_cursor": False,
             "states": per_partition_state,
         }
+
+
+@dataclass
+class CustomReportExtractDimensionsFromKeys(RecordTransformation):
+    """
+    A record transformation that remaps each value in the keys array back to its associated
+    dimension. The reason this is a custom component is because we're unable to use list
+    comprehension and and enumerate() is not a valid function in our Jinja contact so can't
+    iterate over the dimensions defined in the config to create each field transformation on the
+    stream_template for each custom report.
+
+    If we were able to, the actual ComponentMappingDefinition would look like this:
+
+    type: ComponentMappingDefinition
+    field_path:
+      - transformations
+      - "1"
+      - fields
+    value: "{{ [{'path': [dimension], 'value': '{{ record['keys'][index]} for index, dimension in enumerate(components_values['dimensions'])] }}"
+
+    or
+
+    type: ComponentMappingDefinition
+    field_path:
+      - transformations
+      - "1"
+      - fields
+    value: >
+      {% for index, dimension in enumerate(components_values["dimensions"]) %}
+           - type: AddFields
+             fields:
+               - path: [ {{ dimension }} ]
+                 value: "{{ record['keys'][index] }}"
+      {% endfor %}
+    """
+
+    dimensions: List[str] = field(default_factory=lambda: [])
+
+    def transform(
+        self,
+        record: Dict[str, Any],
+        config: Optional[Config] = None,
+        stream_state: Optional[StreamState] = None,
+        stream_slice: Optional[StreamSlice] = None,
+    ) -> None:
+        for dimension in self.dimensions:
+            record[dimension] = record["keys"].pop(0)
+
+        record.pop("keys")
+
+
+@dataclass
+class CustomReportSchemaLoader(SchemaLoader):
+    """
+    Custom schema loader is needed because Google Search Console's custom reports streams
+    because the schema is dependent on which dimensions are selected in the config. Right now,
+    only DynamicSchemaLoader which is based on the response from an API endpoint supports
+    remapping of types to Airbyte schema types. This CustomReportSchemaLoader functions
+    more like a static schema loader and so we must perform the remapping in this custom component.
+    """
+
+    DIMENSION_TO_PROPERTY_SCHEMA_MAP = {
+        "country": [{"country": {"type": ["null", "string"]}}],
+        "date": [{"date": {"type": ["null", "string"], "format": "date"}}],
+        "device": [{"device": {"type": ["null", "string"]}}],
+        "page": [{"page": {"type": ["null", "string"]}}],
+        "query": [{"query": {"type": ["null", "string"]}}],
+    }
+
+    dimensions: List[str]
+
+    def get_json_schema(self) -> Mapping[str, Any]:
+        schema: Mapping[str, Any] = {
+            "$schema": "https://json-schema.org/draft-07/schema#",
+            "type": ["null", "object"],
+            "additionalProperties": True,
+            "properties": {
+                # metrics
+                "clicks": {"type": ["null", "integer"]},
+                "ctr": {"type": ["null", "number"], "multipleOf": 1e-25},
+                "impressions": {"type": ["null", "integer"]},
+                "position": {"type": ["null", "number"], "multipleOf": 1e-25},
+                # default fields
+                "search_type": {"type": ["null", "string"]},
+                "site_url": {"type": ["null", "string"]},
+            },
+        }
+
+        # dimensions
+        dimension_properties = self._dimension_to_property_schema()
+        schema["properties"].update(dimension_properties)
+        return schema
+
+    def _dimension_to_property_schema(self) -> dict:
+        properties = {}
+        for dimension in sorted(self.dimensions):
+            fields = self.DIMENSION_TO_PROPERTY_SCHEMA_MAP[dimension]
+            for field in fields:
+                properties = {**properties, **field}
+        return properties