Commit bd5dc5e2 authored by Jan Reimes's avatar Jan Reimes
Browse files

feat(02-03): Consolidate load_dotenv calls in CLI 

- Move load_dotenv() to tdoc_app.py at module level
- Remove load_dotenv() from crawl.py, query.py, specs.py
- Add noqa comments for E402 (imports after load_dotenv)
- Fix __all__ export in tdoc_app.py (app -> tdoc_app)
- Add comment explaining why load_dotenv is at module level
- Create tests/test_cli_dotenv.py to verify consolidation

All CLI submodules now rely on tdoc_app.py loading .env first.
parent 5798d900

src/tdoc_crawler/cli/crawl.py

+0 −3
Original line number Diff line number Diff line
@@ -6,7 +6,6 @@ import asyncio 
from datetime import date, datetime



import typer

from dotenv import load_dotenv



from tdoc_crawler.cli._shared import console, create_progress_bar, handle_clear_options

from tdoc_crawler.cli.args import (
@@ -67,8 +66,6 @@ from tdoc_crawler.tdocs.operations.crawl import TDocCrawlResult 
from tdoc_crawler.utils.date_parser import parse_partial_date

from tdoc_crawler.utils.parse import collect_spec_numbers, parse_subgroups, parse_working_groups



load_dotenv()



HELP_PANEL = "Crawling Commands"

src/tdoc_crawler/cli/query.py

+0 −3
Original line number Diff line number Diff line
@@ -7,7 +7,6 @@ from datetime import UTC, datetime 
from typing import Annotated



import typer

from dotenv import load_dotenv



from tdoc_crawler.cli._shared import console, handle_clear_options

from tdoc_crawler.cli.args import (
@@ -62,8 +61,6 @@ from tdoc_crawler.tdocs.operations.fetch import fetch_missing_tdocs 
from tdoc_crawler.utils.date_parser import parse_partial_date

from tdoc_crawler.utils.parse import collect_spec_numbers, parse_subgroups, parse_working_groups



load_dotenv()



HELP_PANEL = "Query Commands"

src/tdoc_crawler/cli/specs.py

+0 −3
Original line number Diff line number Diff line
@@ -7,7 +7,6 @@ from pathlib import Path 
from typing import Annotated



import typer

from dotenv import load_dotenv



from tdoc_crawler.cli._shared import console

from tdoc_crawler.cli.args import (
@@ -29,8 +28,6 @@ from tdoc_crawler.specs.downloads import SpecDownloads 
from tdoc_crawler.specs.operations.checkout import build_default_spec_sources

from tdoc_crawler.utils.parse import collect_spec_numbers



load_dotenv()



HELP_PANEL = "Main Commands"

src/tdoc_crawler/cli/tdoc_app.py

+28 −24
Original line number Diff line number Diff line
@@ -2,16 +2,22 @@ 


from __future__ import annotations



import asyncio

import zipfile

from typing import Any, cast



import typer

# Load environment variables from .env file before any other imports

# that might read from environment. This ensures .env values are available

# for TDocCrawlerConfig and other components.

from dotenv import load_dotenv

from rich.table import Table



from tdoc_crawler.cli._shared import console, create_progress_bar

from tdoc_crawler.cli.args import (

load_dotenv()



import asyncio  # noqa: E402

import zipfile  # noqa: E402

from typing import Any, cast  # noqa: E402



import typer  # noqa: E402

from rich.table import Table  # noqa: E402



from tdoc_crawler.cli._shared import console, create_progress_bar  # noqa: E402

from tdoc_crawler.cli.args import (  # noqa: E402

    CacheDirOption,

    CheckoutTDocIdsArgument,

    EolPasswordOption,
@@ -23,21 +29,19 @@ from tdoc_crawler.cli.args import ( 
    UseWhatTheSpecOption,

    VerbosityOption,

)

from tdoc_crawler.cli.crawl import crawl_meetings, crawl_tdocs

from tdoc_crawler.cli.query import query_meetings, query_tdocs

from tdoc_crawler.cli.utils import launch_file

from tdoc_crawler.config import CacheManager

from tdoc_crawler.credentials import set_credentials

from tdoc_crawler.database import MeetingDatabase, TDocDatabase

from tdoc_crawler.http_client import create_cached_session

from tdoc_crawler.logging import DEFAULT_LEVEL as DEFAULT_VERBOSITY

from tdoc_crawler.logging import set_verbosity

from tdoc_crawler.tdocs.models import TDocQueryConfig

from tdoc_crawler.tdocs.operations.checkout import checkout_tdoc, prepare_tdoc_file

from tdoc_crawler.tdocs.operations.fetch import fetch_missing_tdocs

from tdoc_crawler.utils.normalization import normalize_tdoc_id, normalize_tdoc_ids



load_dotenv()

from tdoc_crawler.cli.crawl import crawl_meetings, crawl_tdocs  # noqa: E402

from tdoc_crawler.cli.query import query_meetings, query_tdocs  # noqa: E402

from tdoc_crawler.cli.utils import launch_file  # noqa: E402

from tdoc_crawler.config import CacheManager  # noqa: E402

from tdoc_crawler.credentials import set_credentials  # noqa: E402

from tdoc_crawler.database import MeetingDatabase, TDocDatabase  # noqa: E402

from tdoc_crawler.http_client import create_cached_session  # noqa: E402

from tdoc_crawler.logging import DEFAULT_LEVEL as DEFAULT_VERBOSITY  # noqa: E402

from tdoc_crawler.logging import set_verbosity  # noqa: E402

from tdoc_crawler.tdocs.models import TDocQueryConfig  # noqa: E402

from tdoc_crawler.tdocs.operations.checkout import checkout_tdoc, prepare_tdoc_file  # noqa: E402

from tdoc_crawler.tdocs.operations.fetch import fetch_missing_tdocs  # noqa: E402

from tdoc_crawler.utils.normalization import normalize_tdoc_id, normalize_tdoc_ids  # noqa: E402



tdoc_app = typer.Typer(help="3GPP Crawler - TDocs and Meetings")
@@ -268,4 +272,4 @@ tdoc_app.command("qt", rich_help_panel=HELP_PANEL_QUERY, hidden=True)(query_tdoc 
tdoc_app.command("qm", rich_help_panel=HELP_PANEL_QUERY, hidden=True)(query_meetings)





__all__ = ["app"]
 
__all__ = ["tdoc_app"]

tests/test_cli_dotenv.py

0 → 100644
+102 −0
Original line number Diff line number Diff line 
"""Tests for load_dotenv consolidation in CLI entry point."""



from __future__ import annotations



import ast

from pathlib import Path



import pytest





class TestLoadDotenvConsolidation:

    """Test that load_dotenv is called once at CLI entry."""



    def test_no_duplicate_dotenv_calls(self) -> None:

        """Verify load_dotenv is not called in submodule CLI files."""

        cli_dir = Path(__file__).parent.parent / "src" / "tdoc_crawler" / "cli"

        submodule_files = [

            "crawl.py",

            "query.py",

            "specs.py",

        ]



        for filename in submodule_files:

            filepath = cli_dir / filename

            if filepath.exists():

                source = filepath.read_text()

                tree = ast.parse(source)



                # Check for load_dotenv import

                has_import = any(

                    isinstance(node, ast.ImportFrom) and node.module == "dotenv" and any(alias.name == "load_dotenv" for alias in node.names)

                    for node in ast.walk(tree)

                )



                # Check for load_dotenv() call

                has_call = any(isinstance(node, ast.Call) and isinstance(node.func, ast.Name) and node.func.id == "load_dotenv" for node in ast.walk(tree))



                assert not has_import, f"{filename} should not import load_dotenv"

                assert not has_call, f"{filename} should not call load_dotenv()"



    def test_tdoc_app_has_load_dotenv(self) -> None:

        """tdoc_app.py should import and call load_dotenv."""

        app_file = Path(__file__).parent.parent / "src" / "tdoc_crawler" / "cli" / "tdoc_app.py"

        source = app_file.read_text()

        tree = ast.parse(source)



        # Check for load_dotenv import

        has_import = any(

            isinstance(node, ast.ImportFrom) and node.module == "dotenv" and any(alias.name == "load_dotenv" for alias in node.names) for node in ast.walk(tree)

        )



        # Check for load_dotenv() call

        has_call = any(isinstance(node, ast.Call) and isinstance(node.func, ast.Name) and node.func.id == "load_dotenv" for node in ast.walk(tree))



        assert has_import, "tdoc_app.py should import load_dotenv"

        assert has_call, "tdoc_app.py should call load_dotenv()"





class TestBackwardCompatibility:

    """Test that existing .env patterns still work."""



    def test_existing_env_vars_still_work(self, monkeypatch) -> None:

        """Direct environment variables still take precedence."""

        test_path = Path.home() / ".explicit_cache_test"

        monkeypatch.setenv("TDC_CACHE_DIR", str(test_path))



        from tdoc_crawler.config import TDocCrawlerConfig



        config = TDocCrawlerConfig()



        assert config.path.cache_dir == test_path



    def test_http_cache_env_vars_work(self, monkeypatch) -> None:

        """HTTP_CACHE_TTL env var still works."""

        monkeypatch.setenv("HTTP_CACHE_TTL", "3600")



        from tdoc_crawler.config import TDocCrawlerConfig



        config = TDocCrawlerConfig()



        assert config.http.cache_ttl == 3600



    def test_dotenv_called_before_other_imports(self) -> None:

        """load_dotenv() is called at module level before other imports."""

        app_file = Path(__file__).parent.parent / "src" / "tdoc_crawler" / "cli" / "tdoc_app.py"

        source = app_file.read_text()



        # Find line numbers of key elements

        lines = source.split("\n")



        load_dotenv_line = -1

        typer_import_line = -1



        for i, line in enumerate(lines):

            if "load_dotenv()" in line and not line.strip().startswith("#"):

                load_dotenv_line = i

            if "import typer" in line and not line.strip().startswith("#"):

                typer_import_line = i



        assert load_dotenv_line > 0, "load_dotenv() call should exist"

        assert typer_import_line > 0, "import typer should exist"

        assert load_dotenv_line < typer_import_line, "load_dotenv() should be called before other imports"