[sc-27944] Add config to disable tag collection in Snowflake (#992)

metaphor/snowflake/README.md

@@ -122,6 +122,14 @@ account_usage_schema: <db_name>.<schema_name>
 
 See [Tag Matcher Config](../common/docs/tag_matcher.md) for more information on the optional `tag_matcher` config.
 
+#### Disable Platform Tags Collection
+
+To stop the crawler from collecting platform tags from Snowflake, set `collect_tags` to `False`:
+
+```yaml
+collect_tags: false # Default is true.
+```
+
 #### Query Logs
 
 By default, the snowflake connector will fetch a full day's query logs from yesterday, to be analyzed for additional metadata, such as dataset usage and lineage information. To backfill log data, one can set `lookback_days` to the desired value. To turn off query log fetching, set `lookback_days` to 0.  

metaphor/snowflake/config.py

@@ -58,6 +58,9 @@ class SnowflakeBaseConfig(SnowflakeAuthConfig):
     # The fully qualified schema that contains all the account_usage views
     account_usage_schema: str = "SNOWFLAKE.ACCOUNT_USAGE"
 
+    # Whether to collect platform tags.
+    collect_tags: bool = True
+
 
 @dataclass(config=ConnectorConfig)
 class SnowflakeConfig(SnowflakeBaseConfig):

metaphor/snowflake/extractor.py

@@ -171,7 +171,10 @@ async def extract(self) -> Collection[ENTITY_TYPES]:
 
             self._fetch_primary_keys(cursor)
             self._fetch_unique_keys(cursor)
-            self._fetch_tags(cursor)
+
+            # Only fetch the tags when collect_tags is True
+            if self._config.collect_tags:
+                self._fetch_tags(cursor)
 
         datasets = list(self._datasets.values())
         tag_datasets(datasets, self._tag_matchers)
@@ -902,7 +905,9 @@ def _init_dataset(
             database=database, schema=schema, table=table
         )
 
-        dataset.system_tags = SystemTags(tags=[])
+        # Only initialize this when collect_tags is True
+        if self._config.collect_tags:
+            dataset.system_tags = SystemTags(tags=[])
 
         return dataset
 

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "metaphor-connectors"
-version = "0.14.110"
+version = "0.14.111"
 license = "Apache-2.0"
 description = "A collection of Python-based 'connectors' that extract metadata from various sources to ingest into the Metaphor app."
 authors = ["Metaphor <[email protected]>"]