Commit 5cd33bef authored by Jan Reimes's avatar Jan Reimes
Browse files

feat(cli): add no-progress option to disable progress bar during crawl 

* Introduced NoProgressOption to allow users to disable the progress bar.
* Updated crawl_tdocs function to handle no progress scenario.
* Adjusted progress handling logic based on the no_progress flag.
parent f95f9d2f

src/tdoc_crawler/cli/args.py

+7 −0
Original line number Diff line number Diff line
@@ -23,9 +23,11 @@ SubgroupOption = Annotated[ 
    typer.Option("--sub-group", "-s", help="Filter by sub-working group (e.g., 'R2-102')", envvar="TDC_SUB_GROUP"),

]

LimitMeetingsOption = Annotated[int | None, typer.Option("--limit-meetings", help="Limit meetings overall", envvar="TDC_LIMIT_MEETINGS")]

# TODO: This parameter does not make sense for working groups, but would for sub-workinggroups. Consider renaming to --limit-meetings-per-swg.

LimitMeetingsPerWgOption = Annotated[

    int | None, typer.Option("--limit-meetings-per-wg", help="Limit meetings per working group", envvar="TDC_LIMIT_MEETINGS_PER_WG")

]

# TODO: This parameter does not make sense for working groups, but would for sub-workinggroups.

LimitWgsOption = Annotated[int | None, typer.Option("--limit-wgs", help="Limit number of working groups")]

LimitOption = Annotated[int | None, typer.Option("--limit", "-l", help="Maximum number of rows")]

OrderOption = Annotated[str, typer.Option("--order", help="Sort order (asc|desc)")]
@@ -118,3 +120,8 @@ HttpCacheOption = Annotated[ 
        envvar="HTTP_CACHE_ENABLED",

    ),

]



NoProgressOption = Annotated[

    bool,

    typer.Option("--no-progress", help="Disable progress bar (useful for scripts and CI)"),

]

src/tdoc_crawler/cli/crawl.py

+11 −5
Original line number Diff line number Diff line
@@ -26,6 +26,7 @@ from tdoc_crawler.cli.args import ( 
    LimitTDocsOption,

    LimitWgsOption,

    MaxRetriesOption,

    NoProgressOption,

    OutputFormatOption,

    OverallTimeoutOption,

    PromptCredentialsOption,
@@ -76,6 +77,7 @@ def crawl_tdocs( 
    timeout: TimeoutOption = 30,

    max_retries: MaxRetriesOption = 3,

    overall_timeout: OverallTimeoutOption = None,

    no_progress: NoProgressOption = False,

    cache_dir: CacheDirOption = None,

    http_cache_enabled: HttpCacheOption = None,

    verbosity: VerbosityOption = str(DEFAULT_VERBOSITY),
@@ -133,6 +135,10 @@ def crawl_tdocs( 


        crawl_start_time = datetime.now()



        if no_progress:

            # No progress bar - just run the crawl

            result = crawler.crawl(config, progress_callback=None)

        else:

            progress, task = create_progress_bar("[cyan]Crawling TDocs...")



            with progress: