chore(ingest): cleanup various methods (#9221)

metadata-events/mxe-schemas/rename-namespace.sh

@@ -1,4 +1,4 @@
-#!/bin/sh
+#!/bin/bash
 
 SCRIPT_ROOT="$( cd "$( dirname "${BASH_SOURCE[0]:-$0}" )" >/dev/null && pwd )"
 

metadata-ingestion/scripts/docgen.py

@@ -7,11 +7,10 @@ import re
 import sys
 import textwrap
 from importlib.metadata import metadata, requires
-from typing import Any, Dict, Iterable, List, Optional, Tuple, Union
+from typing import Any, Dict, Iterable, List, Optional
 
 import click
 from pydantic import BaseModel, Field
-from pydantic.dataclasses import dataclass
 
 from datahub.configuration.common import ConfigModel
 from datahub.ingestion.api.decorators import (
@@ -94,7 +93,6 @@ class FieldRow(BaseModel):
 
     @staticmethod
     def map_field_path_to_components(field_path: str) -> List[Component]:
-
         m = re.match(FieldRow._V2_FIELD_PATH_TOKEN_MATCHER_PREFIX, field_path)
         v = re.match(FieldRow._V2_FIELD_PATH_FIELD_NAME_MATCHER, field_path)
         components: List[FieldRow.Component] = []
@@ -197,7 +195,7 @@ class FieldRow(BaseModel):
             # Using a non-breaking space to prevent the checkbox from being
             # broken into a new line.
             if not self.parent:  # None and empty string both count
-                return f'&nbsp;<abbr title="Required">✅</abbr>'
+                return '&nbsp;<abbr title="Required">✅</abbr>'
             else:
                 return f'&nbsp;<abbr title="Required if {self.parent} is set">❓</abbr>'
         else:
@@ -356,7 +354,6 @@ def priority_value(path: str) -> str:
 
 
 def gen_md_table_from_struct(schema_dict: Dict[str, Any]) -> List[str]:
-
     from datahub.ingestion.extractor.json_schema_util import JsonSchemaTranslator
 
     # we don't want default field values to be injected into the description of the field
@@ -460,7 +457,6 @@ def get_additional_deps_for_extra(extra_name: str) -> List[str]:
 
 
 def relocate_path(orig_path: str, relative_path: str, relocated_path: str) -> str:
-
     newPath = os.path.join(os.path.dirname(orig_path), relative_path)
     assert os.path.exists(newPath)
 
@@ -515,7 +511,6 @@ def generate(
 
     if extra_docs:
         for path in glob.glob(f"{extra_docs}/**/*[.md|.yaml|.yml]", recursive=True):
-
             m = re.search("/docs/sources/(.*)/(.*).md", path)
             if m:
                 platform_name = m.group(1).lower()
@@ -741,7 +736,7 @@ def generate(
             i += 1
             f.write(f"---\nsidebar_position: {i}\n---\n\n")
             f.write(
-                f"import Tabs from '@theme/Tabs';\nimport TabItem from '@theme/TabItem';\n\n"
+                "import Tabs from '@theme/Tabs';\nimport TabItem from '@theme/TabItem';\n\n"
             )
             f.write(f"# {platform_docs['name']}\n")
 

metadata-ingestion/src/datahub/ingestion/source/looker/looker_common.py

@@ -828,7 +828,7 @@ class LookerExplore:
                 )
             else:
                 logger.warning(
-                    f"Failed to extract explore {explore_name} from model {model}.", e
+                    f"Failed to extract explore {explore_name} from model {model}: {e}"
                 )
 
         except AssertionError:

metadata-ingestion/src/datahub/ingestion/source/mode.py

@@ -218,6 +218,8 @@ class ModeSource(Source):
         if creator is not None:
             modified_actor = builder.make_user_urn(creator)
             if report_info.get("last_saved_at") is None:
+                # Sometimes mode returns null for last_saved_at.
+                # In that case, we use the created_at timestamp instead.
                 report_info["last_saved_at"] = report_info.get("created_at")
 
             modified_ts = int(