koko210Serve fc4674bb13 refactor: extract media processing from bot.py into image_handling.py (Phase D Step 19) ...

- Create process_media_in_message() in utils/image_handling.py that handles all 4 media
  types: image attachments, video/GIF attachments, Tenor GIF embeds, and rich embeds
- DRY the send→log→bipolar tail pattern (5x repeated) into _send_log_bipolar() helper
- Unify rich/article/link embed handling to use rephrase_as_miku() instead of inline
  Cat→LLM routing, fixing a mood-resolution bug (was using globals.DM_MOOD for servers)
- Add 'rich_embed' media_type to rephrase_as_miku() prefix switch
- Remove 3 inline 'import base64' from bot.py (already module-level in image_handling.py)
- bot.py: 986 → 623 lines (-363)
- image_handling.py: 559 → 881 lines (+322)
- All 170 tests pass (21 config/state + 149 route split)

2026-04-15 12:19:37 +03:00

35 KiB

Raw Blame History

View Raw