refactor(core): rename shadowing variables and add meeting helper (fc06dce3) · Commits · Jan Reimes / 3gpp-crawler

src/tdoc_crawler/cli/config_app.py

+5 −5

Original line number	Diff line number	Diff line
		@@ -80,12 +80,12 @@ def config_show(
		print(exporter.export(format))


		def _check_path_exists(path: Path) -> tuple[bool, str]:
		"""Check if a path exists or can be created."""
		if path.exists():
		def _check_dir_exists(target_dir: Path) -> tuple[bool, str]:
		"""Check if a directory exists or can be created."""
		if target_dir.exists():
		return True, ""
		try:
		path.parent.mkdir(parents=True, exist_ok=True)
		target_dir.parent.mkdir(parents=True, exist_ok=True)
		return True, " (will be created)"
		except PermissionError:
		return False, " (permission denied)"
		@@ -99,7 +99,7 @@ def _validate_config_values(config: ThreeGPPConfig) -> list[tuple[str, str]]:

		# Check cache_dir
		cache_dir = config.path.cache_dir
		exists, detail = _check_path_exists(cache_dir)
		exists, detail = _check_dir_exists(cache_dir)
		if not exists:
		issues.append(("error", f"cache_dir does not exist and cannot be created: {cache_dir}{detail}"))
		elif detail:

src/tdoc_crawler/cli/query.py

+12 −12

Original line number	Diff line number	Diff line
		@@ -89,7 +89,7 @@ def query_tdocs(
		) -> None:
		"""Query TDoc metadata from database."""
		set_verbosity(verbosity)
		path = PathConfig(cache_dir=cache_dir) if cache_dir else PathConfig()
		path_config = PathConfig(cache_dir=cache_dir) if cache_dir else PathConfig()

		working_groups = parse_working_groups(working_group)
		try:
		@@ -131,10 +131,10 @@ def query_tdocs(
		agenda_pattern_exclude=agenda_ex,
		)

		db_file = path.db_file
		db_file = path_config.db_file
		handle_clear_options(
		db_file,
		path.checkout_dir,
		path_config.checkout_dir,
		TDocDatabase,
		clear_tdocs=clear_tdocs,
		clear_specs=clear_specs,
		@@ -168,7 +168,7 @@ def query_tdocs(

		if checkout:
		with create_cached_session() as session:
		checkout_tdocs(results, path.checkout_dir, force=False, session=session)
		checkout_tdocs(results, path_config.checkout_dir, force=False, session=session)

		# Build meeting map for enriched output
		async def load_meeting_map() -> dict:
		@@ -206,7 +206,7 @@ def query_meetings(
		) -> None:
		"""Query meeting metadata from database."""
		set_verbosity(verbosity)
		path = PathConfig(cache_dir=cache_dir) if cache_dir else PathConfig()
		path_config = PathConfig(cache_dir=cache_dir) if cache_dir else PathConfig()
		working_groups = parse_working_groups(working_group)
		subgroups = parse_subgroups(subgroup)
		try:
		@@ -223,10 +223,10 @@ def query_meetings(
		include_without_files=include_without_files,
		)

		db_file = path.db_file
		db_file = path_config.db_file
		handle_clear_options(
		db_file,
		path.checkout_dir,
		path_config.checkout_dir,
		MeetingDatabase,
		clear_tdocs=clear_tdocs,
		clear_specs=clear_specs,
		@@ -244,7 +244,7 @@ def query_meetings(

		if checkout:
		with create_cached_session() as session:
		checkout_meeting_tdocs(meetings, path.checkout_dir, path.http_cache_file, session=session)
		checkout_meeting_tdocs(meetings, path_config.checkout_dir, path_config.http_cache_file, session=session)

		try:
		output = OutputFormat(output_format.lower())
		@@ -275,7 +275,7 @@ def query_specs(
		) -> None:
		"""Query spec metadata from database."""
		set_verbosity(verbosity)
		path = PathConfig(cache_dir=cache_dir) if cache_dir else PathConfig()
		path_config = PathConfig(cache_dir=cache_dir) if cache_dir else PathConfig()
		specs = collect_spec_numbers(spec_numbers, spec_file)
		working_groups = parse_working_groups(working_group)
		wg_filter = working_groups[0].value if working_groups else None
		@@ -293,10 +293,10 @@ def query_specs(
		console.print("[red]Invalid output format; use table, json, jsonl, toon, or yaml")
		raise typer.Exit(code=2) from exc

		db_file = path.db_file
		db_file = path_config.db_file
		handle_clear_options(
		db_file,
		path.checkout_dir,
		path_config.checkout_dir,
		SpecDatabase,
		clear_tdocs=clear_tdocs,
		clear_specs=clear_specs,
		@@ -317,7 +317,7 @@ def query_specs(

		async def load_specs_for_checkout() -> None:
		async with SpecDatabase(db_file) as database:
		checkout_specs(spec_list, path.checkout_dir, database, release="latest")
		checkout_specs(spec_list, path_config.checkout_dir, database, release="latest")

		asyncio.run(load_specs_for_checkout())

src/tdoc_crawler/config/sources.py

+17 −17

Original line number	Diff line number	Diff line
		@@ -40,14 +40,14 @@ class ConfigLoadError(Exception):
		"""Raised when a config file cannot be loaded or parsed.

		Attributes:
		file_path: Path to the file that failed to load.
		config_file: Path to the file that failed to load.
		reason: Human-readable explanation of the failure.
		"""

		def __init__(self, file_path: Path, reason: str) -> None:
		self.file_path = file_path
		def __init__(self, config_file: Path, reason: str) -> None:
		self.config_file = config_file
		self.reason = reason
		super().__init__(f"Failed to load config from {file_path}: {reason}")
		super().__init__(f"Failed to load config from {config_file}: {reason}")


		def _interpolate_env_vars(value: Any) -> Any:
		@@ -145,14 +145,14 @@ def discover_config_files(cwd: Path \| None = None) -> list[Path]:
		return files


		def load_config_file(path: Path) -> dict[str, Any]:
		def load_config_file(config_file: Path) -> dict[str, Any]:
		"""Load a single config file and return its contents as a dict.

		Supports TOML (.toml), YAML (.yaml, .yml), and JSON (.json) formats.
		Environment variable interpolation is applied to string values.

		Args:
		path: Path to the config file.
		config_file: Path to the config file.

		Returns:
		Dictionary of config values. Empty dict if file not found.
		@@ -160,49 +160,49 @@ def load_config_file(path: Path) -> dict[str, Any]:
		Raises:
		ConfigLoadError: If the file cannot be parsed or has permission issues.
		"""
		if not path.is_file():
		if not config_file.is_file():
		return {}

		suffix = path.suffix.lower()
		suffix = config_file.suffix.lower()

		try:
		if suffix == ".toml":
		# For TOML, we need to interpolate BEFORE parsing since TOML
		# doesn't support ${VAR} syntax natively
		with path.open("r", encoding="utf-8") as f:
		with config_file.open("r", encoding="utf-8") as f:
		content = f.read()
		content = _interpolate_env_vars(content)
		data = tomllib.loads(content)
		elif suffix in {".yaml", ".yml"}:
		yaml = import_module("yaml")

		with path.open("r", encoding="utf-8") as f:
		with config_file.open("r", encoding="utf-8") as f:
		try:
		data = yaml.safe_load(f)
		except yaml.scanner.ScannerError as e:
		raise ConfigLoadError(path, f"Parse error: {e}") from e
		raise ConfigLoadError(config_file, f"Parse error: {e}") from e
		if data is None:
		data = {}
		data = _interpolate_env_vars(data)
		elif suffix == ".json":
		with path.open("r", encoding="utf-8") as f:
		with config_file.open("r", encoding="utf-8") as f:
		data = json.load(f)
		data = _interpolate_env_vars(data)
		else:
		raise ConfigLoadError(path, f"Unsupported file format: {suffix}")
		raise ConfigLoadError(config_file, f"Unsupported file format: {suffix}")
		except FileNotFoundError:
		return {}
		except PermissionError as e:
		raise ConfigLoadError(path, f"Permission denied: {e}") from e
		raise ConfigLoadError(config_file, f"Permission denied: {e}") from e
		except (ValueError, json.JSONDecodeError) as e:
		raise ConfigLoadError(path, f"Parse error: {e}") from e
		raise ConfigLoadError(config_file, f"Parse error: {e}") from e
		except ImportError as e:
		if "yaml" in str(e):
		raise ConfigLoadError(path, "PyYAML is required for YAML files but is not installed") from e
		raise ConfigLoadError(config_file, "PyYAML is required for YAML files but is not installed") from e
		raise

		if not isinstance(data, dict):
		raise ConfigLoadError(path, f"Expected dict at root, got {type(data).__name__}")
		raise ConfigLoadError(config_file, f"Expected dict at root, got {type(data).__name__}")

		return data

src/tdoc_crawler/database/meetings.py

+31 −1

Original line number	Diff line number	Diff line
		"""Meeting database operations."""

		import asyncio
		from collections import defaultdict
		from collections.abc import Callable, Iterable
		from datetime import datetime

		from tdoc_crawler.config.settings import PathConfig
		from tdoc_crawler.database.base import DocDatabase
		from tdoc_crawler.database.oxyde_models import CrawlLogEntry, MeetingMetadata, TDocMetadata
		from tdoc_crawler.logging import get_logger
		@@ -294,4 +296,32 @@ class MeetingDatabase(DocDatabase):
		return False


		__all__ = ["MeetingDatabase"]
		async def _resolve_meeting_short_name(meeting_id: int) -> str \| None:
		"""Look up a meeting's short name by ID.

		Args:
		meeting_id: Meeting identifier.

		Returns:
		Meeting short name (e.g., "SA4#134") if found, None otherwise.
		"""
		async with MeetingDatabase(PathConfig().db_file) as db:
		meeting = await db._get_meeting(meeting_id)
		return meeting.short_name if meeting is not None else None


		def get_meeting_short_name(meeting_id: int) -> str \| None:
		"""Sync convenience wrapper to look up a meeting's short name by ID.

		Uses ``asyncio.run()`` — must not be called from an async context.

		Args:
		meeting_id: Meeting identifier.

		Returns:
		Meeting short name (e.g., "SA4#134") if found, None otherwise.
		"""
		return asyncio.run(_resolve_meeting_short_name(meeting_id))


		__all__ = ["MeetingDatabase", "get_meeting_short_name"]

tests/test_targeted_fetch.py

+2 −2

Original line number	Diff line number	Diff line
		@@ -60,10 +60,10 @@ class TestInferWorkingGroups:
		assert WorkingGroup.RAN in groups

		def test_infer_empty_list(self) -> None:
		"""Return all working groups for empty input."""
		"""Return empty list for empty input."""
		ids: list[str] = []
		groups = infer_working_groups_from_ids(ids)
		assert len(groups) == 3
		assert len(groups) == 0


		@pytest.mark.asyncio