Commit 94d55147 authored by Jan Reimes's avatar Jan Reimes
Browse files

feat(03-02): update tdoc_app.py with --config support 

- Add --config option to callback with typer.Option
- Remove module-level load_dotenv() - consolidated at entry point
- Update open_tdoc, checkout, stats commands to use ctx.obj config
- Add deprecation warning for --cache-dir option
- Commands now get CacheManager from config instead of explicit cache_dir
parent 7cd924ed

src/tdoc_crawler/cli/tdoc_app.py

+60 −35
Original line number Diff line number Diff line
@@ -2,22 +2,16 @@ 


from __future__ import annotations



# 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

import asyncio

import zipfile

from pathlib import Path

from typing import Annotated, Any, cast



load_dotenv()

import typer

from rich.table import Table



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

from tdoc_crawler.cli._shared import console, create_progress_bar

from tdoc_crawler.cli.args import (

    CacheDirOption,

    CheckoutTDocIdsArgument,

    EolPasswordOption,
@@ -29,19 +23,20 @@ from tdoc_crawler.cli.args import ( # noqa: E402 
    UseWhatTheSpecOption,

    VerbosityOption,

)

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

from tdoc_crawler.cli.config import load_cli_config

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



tdoc_app = typer.Typer(help="3GPP Crawler - TDocs and Meetings")
@@ -56,10 +51,34 @@ HELP_PANEL_QUERY = "Query Commands" 
@tdoc_app.callback()

def _app_callback(

    ctx: typer.Context,

    cache_dir: CacheDirOption = None,

    config_file: Annotated[

        Path | None,

        typer.Option(

            "--config",

            "-c",

            help="Path to configuration file (overrides discovered config)",

            exists=True,

            readable=True,

        ),

    ] = None,

    cache_dir: Annotated[

        Path | None,

        typer.Option(

            "--cache-dir",

            help="Cache directory (deprecated: use config file)",

            envvar="TDC_CACHE_DIR",

        ),

    ] = None,

) -> None:

    """Global CLI options."""

    pass  # No global options currently

    # Load and validate config at startup

    config = load_cli_config(ctx, config_file)



    # For backward compatibility: if --cache-dir provided, it overrides config

    if cache_dir:

        console.print("[yellow]Warning: --cache-dir is deprecated, use config file[/yellow]")

        # Create new CacheManager with explicit cache_dir

        CacheManager(cache_dir).register()





# Register crawl commands
@@ -74,13 +93,13 @@ tdoc_app.command("query-meetings", rich_help_panel=HELP_PANEL_QUERY)(query_meeti 


@tdoc_app.command("open", rich_help_panel=HELP_PANEL_MAIN)

def open_tdoc(

    ctx: typer.Context,

    tdoc_id: TDocIdArgument,

    full_metadata: FullMetadataOption = False,

    use_whatthespec: UseWhatTheSpecOption = False,

    eol_username: EolUsernameOption = None,

    eol_password: EolPasswordOption = None,

    prompt_credentials: PromptCredentialsOption = None,

    cache_dir: CacheDirOption = None,

    verbosity: VerbosityOption = str(DEFAULT_VERBOSITY),

) -> None:

    """Download, extract, and open a TDoc file."""
@@ -88,7 +107,9 @@ def open_tdoc( 


    set_credentials(eol_username, eol_password, prompt=prompt_credentials)



    manager = CacheManager(cache_dir).register()

    # Get config from context (set in callback)

    crawler_config = ctx.obj

    manager = CacheManager(config=crawler_config).register()



    normalized_id = normalize_tdoc_id(tdoc_id)
@@ -139,6 +160,7 @@ def open_tdoc( 


@tdoc_app.command("checkout", rich_help_panel=HELP_PANEL_MAIN)

def checkout(

    ctx: typer.Context,

    tdoc_ids: CheckoutTDocIdsArgument,

    force: ForceOption = False,

    full_metadata: FullMetadataOption = False,
@@ -146,7 +168,6 @@ def checkout( 
    eol_username: EolUsernameOption = None,

    eol_password: EolPasswordOption = None,

    prompt_credentials: PromptCredentialsOption = None,

    cache_dir: CacheDirOption = None,

    verbosity: VerbosityOption = str(DEFAULT_VERBOSITY),

) -> None:

    """Download and extract TDoc(s) to checkout folder."""
@@ -154,7 +175,9 @@ def checkout( 


    set_credentials(eol_username, eol_password, prompt=prompt_credentials)



    manager = CacheManager(cache_dir).register()

    # Get config from context (set in callback)

    crawler_config = ctx.obj

    manager = CacheManager(config=crawler_config).register()



    normalized_ids = normalize_tdoc_ids(tdoc_ids)
@@ -220,13 +243,15 @@ def checkout( 


@tdoc_app.command("stats", rich_help_panel=HELP_PANEL_MAIN)

def stats(

    cache_dir: CacheDirOption = None,

    ctx: typer.Context,

    verbosity: VerbosityOption = str(DEFAULT_VERBOSITY),

) -> None:

    """Display database statistics."""

    set_verbosity(verbosity)



    manager = CacheManager(cache_dir).register()

    # Get config from context (set in callback)

    crawler_config = ctx.obj

    manager = CacheManager(config=crawler_config).register()



    if not (db_file := manager.db_file).exists():

        console.print(f"[red]Database not found: {db_file}[/red]")