Add optional automatic Speaker Identification to Folder Watch

[mmoroni]

Folder Watch currently automates transcription end-to-end (watch → transcribe → export → optionally delete source), but Speaker Identification / Diarization is only available as a separate, manual post-processing step run from the transcript viewer. For workflows that rely on Folder Watch for hands-off batch processing (e.g. recurring meeting recordings), this breaks the automation — every file still needs to be opened manually to run diarization before the transcript is actually useful.

My suggestion is to add an "Identify Speakers" option to the Folder Watch settings (Preferences → Folder Watch), alongside the existing Model/Task/Language/Export controls. When enabled, Buzz would automatically run speaker identification on each transcript immediately after transcription completes, before export/delete — so the exported TXT/SRT/VTT already contains speaker labels, with no manual step required.

Thank you!

[raivisdejus]

@mmoroni This has been implemented on latest development version