[enh] add multi-language subtitle support to yt-dlp extractor

[dummylabs]

First of all, I really like your project; I use it actively and will follow its development with interest.

Disclaimer

I am not a golang expert; however, I have sufficient experience in other programming languages, so I used Claude Code for AI-assisted development, not vibe-coding. This text was written by a human and verified by an LLM for mistakes.

Problem Statement

If the sub_language setting in the ytdlp extractor is set to a language that does not match the video's default language, the video transcript is not saved in Hister because yt-dlp returns 201 and an empty string.

A workaround is to request an auto-translation of captions into the specified language, but in this case YouTube applies stricter checks, and cookies may be required, overwise it may return 429: Too Many Requests. Therefore, in my experiments, the most reliable approach (without extra impersonation or cookie import) was to request captions in the language that matches the video's default language (info.Language).

Core Changes

In the ytdlp extractor configuration, the sub_language parameter can now take the value auto, or contain one or more subtitle languages separated by commas.
Logic:

• if auto is specified: download subtitles in the default language for this video
• if en is specified: download subtitles in English or nothing
• if en, fr are specified: try downloading subtitles in the listed languages in order of priority, or nothing if none of the specified languages is the video's default.

If en is specified as sub_language, the current behavior (before the change) is fully preserved.

Changes to the default configuration

extractors:
    ytdlp:
        fetch_subtitles: true
        sub_language: "auto"      # detect original language automatically

Testing

Tested manually on videos with different languages and different values of sub_language.

Added new test fixtures:

• sampleJSONTemplateFr for a non-English language test
• sampleJSONTemplateNoLang for a video without a default language (this is the real case in my experience)

Four new tests added:

Test	Fixture	sub_language	Expectation
Auto	Fr	auto	subtitles present
MultiLangMatch	Fr	fr,en	subtitles present
AutoNoLanguage	NoLang	auto	no subtitles
MultiLangNoMatch	Fr	de,es	no subtitles

[asciimoo]

Thank you both for the contribution and for being transparent about AI usage.

The pull requests looks good to me.

I just realized that we do not document extractor specific configuration options anywhere. Would you be open to add yt-dlp config options to the extractor documentation (webui/website/src/content/docs/extractors.md) in a follow-up pull request?

[asciimoo]

Please fix the go lint error and it's ready to merge

[asciimoo]

@dummylabs doh, sorry the lint error is not even in your code.. My fault, I'm fixing it. Not sure how my local linter let that slip..

[asciimoo]

Thanks!