[ninglinLiu]

Related issue

Fixes #13395

Summary

Adds a stdlib-only SubtitleToDocument converter (modeled on TextFileToDocument) that:

• Reads .srt and .vtt
• Strips cue numbers, timing lines, NOTE/STYLE/REGION blocks, and common inline tags
• Groups consecutive cues into Documents via cues_per_document (default 10; None = one Document per file)
• Sets start_time / end_time (float seconds) on each Document from the first/last cue in the group
• Warns and skips unreadable sources or files with no valid cues

MultiFileConverter routing is intentionally unchanged (as discussed in the issue) so existing .srt → TextFileToDocument behavior stays intact.

Verification

pytest test/components/converters/test_subtitle_to_document.py -q
# 6 passed

[vercel]

@ninglinLiu is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[HaystackBot]

Hi @ninglinLiu, thanks a lot for your contribution! 🙏

We noticed that the Contributor License Agreement (CLA) check (license/cla) hasn't passed yet, so we've temporarily moved this PR to draft and paused the review assignment.

To get your PR reviewed, please sign the CLA via the link in the license/cla check below (or in the CLA bot comment). As soon as the check turns green, this PR will automatically be marked ready for review again and a reviewer will be re-assigned.