♻️ refactor(models): normalize working group and subgroup aliases consistently (2794356f) · Commits · Jan Reimes / 3gpp-crawler

src/tdoc_crawler/meetings/models.py

+5 −10

Original line number	Diff line number	Diff line
		@@ -7,6 +7,7 @@ from datetime import date, datetime

		from pydantic import BaseModel, Field, field_validator, model_validator

		from tdoc_crawler.meetings.utils import normalize_subgroup_alias, normalize_working_group_alias
		from tdoc_crawler.models.base import BaseConfigModel, SortOrder
		from tdoc_crawler.models.crawl_limits import CrawlLimits, _new_crawl_limits
		from tdoc_crawler.models.subworking_groups import SUBTB_INDEX
		@@ -80,10 +81,7 @@ class MeetingCrawlConfig(BaseConfigModel):
		@classmethod
		def _normalize_working_groups(cls, value: Iterable[str \| WorkingGroup]) -> list[WorkingGroup]:
		"""Ensure the working groups list only contains valid enum members."""
		normalized: list[WorkingGroup] = []
		for item in value:
		normalized.append(WorkingGroup(item) if not isinstance(item, WorkingGroup) else item)
		return normalized
		return [normalize_working_group_alias(str(item)) if not isinstance(item, WorkingGroup) else item for item in value]

		@field_validator("subgroups", mode="before")
		@classmethod
		@@ -91,7 +89,7 @@ class MeetingCrawlConfig(BaseConfigModel):
		"""Normalize subgroup names (uppercase and strip whitespace)."""
		if value is None:
		return None
		return [str(item).strip().upper() for item in value]
		return [normalize_subgroup_alias(str(item)).value for item in value]


		class MeetingQueryConfig(BaseConfigModel):
		@@ -112,10 +110,7 @@ class MeetingQueryConfig(BaseConfigModel):
		"""Ensure the working group list is comprised of enum members."""
		if value is None:
		return None
		normalized: list[WorkingGroup] = []
		for item in value:
		normalized.append(WorkingGroup(item) if not isinstance(item, WorkingGroup) else item)
		return normalized
		return [normalize_working_group_alias(str(item)) if not isinstance(item, WorkingGroup) else item for item in value]

		@field_validator("subgroups", mode="before")
		@classmethod
		@@ -123,7 +118,7 @@ class MeetingQueryConfig(BaseConfigModel):
		"""Normalize subgroup names (uppercase and strip whitespace)."""
		if value is None:
		return None
		return [str(item).strip().upper() for item in value]
		return [normalize_subgroup_alias(str(item)).value for item in value]


		__all__ = [

src/tdoc_crawler/tdocs/models.py

+6 −11

Original line number	Diff line number	Diff line
		@@ -11,6 +11,7 @@ from packaging.version import Version
		from pydantic import BaseModel, Field, field_serializer, field_validator

		from tdoc_crawler.logging import get_logger
		from tdoc_crawler.meetings.utils import normalize_subgroup_alias, normalize_working_group_alias
		from tdoc_crawler.models.base import (
		BaseConfigModel,
		OutputFormat,
		@@ -18,7 +19,7 @@ from tdoc_crawler.models.base import (
		)
		from tdoc_crawler.models.crawl_limits import CrawlLimits, _new_crawl_limits
		from tdoc_crawler.models.working_groups import WorkingGroup
		from tdoc_crawler.tdocs.utils import normalize_tdoc_ids
		from tdoc_crawler.tdocs.utils import normalize_tdoc_id, normalize_tdoc_ids
		from tdoc_crawler.utils.misc import utc_now
		from tdoc_crawler.utils.parse import AgendaItemNumber, parse_agenda_item_nbr, parse_agenda_item_version

		@@ -220,7 +221,7 @@ class TDocMetadata(BaseModel):
		@classmethod
		def _normalize_tdoc_id(cls, value: str) -> str:
		"""Ensure identifiers are uppercase and trimmed."""
		return value.strip().upper()
		return normalize_tdoc_id(value)

		# Optional metadata fields (from portal or determined otherwise)
		is_revision_of: str \| None = Field(None, description="Reference to a previous TDoc version")
		@@ -273,10 +274,7 @@ class TDocCrawlConfig(BaseConfigModel):
		@classmethod
		def _normalize_working_groups(cls, value: Iterable[str \| WorkingGroup]) -> list[WorkingGroup]:
		"""Ensure the working groups list only contains valid enum members."""
		normalized: list[WorkingGroup] = []
		for item in value:
		normalized.append(WorkingGroup(item) if not isinstance(item, WorkingGroup) else item)
		return normalized
		return [normalize_working_group_alias(str(item)) if not isinstance(item, WorkingGroup) else item for item in value]

		@field_validator("subgroups", mode="before")
		@classmethod
		@@ -284,7 +282,7 @@ class TDocCrawlConfig(BaseConfigModel):
		"""Normalize subgroup names to uppercase."""
		if value is None:
		return None
		return [str(item).upper().strip() for item in value]
		return [normalize_subgroup_alias(str(item)).value for item in value]

		@field_validator("target_ids", mode="before")
		@classmethod
		@@ -329,10 +327,7 @@ class TDocQueryConfig(BaseConfigModel):
		"""Ensure the working group list is comprised of enum members."""
		if value is None:
		return None
		normalized: list[WorkingGroup] = []
		for item in value:
		normalized.append(WorkingGroup(item) if not isinstance(item, WorkingGroup) else item)
		return normalized
		return [normalize_working_group_alias(str(item)) if not isinstance(item, WorkingGroup) else item for item in value]


		__all__ = [