Loi Nguyen
f420e47f95
fix(ollama): preserve multimodal images (#17663)
### Summary
- Convert multimodal text blocks into Ollama message content.
- Attach image references through Ollama native images arrays for
synchronous and streaming chat requests.
- Strip data-URI headers while preserving raw base64 and URL image
values.
- Add regression coverage for both request paths.
Closes #17332
---------
Co-authored-by: Codex <codex@openai.com>
2026-08-20 10:15:15 +08:00
..
2026-08-12 19:11:49 +08:00
2026-08-19 13:47:36 +08:00
2026-08-11 19:36:35 +08:00
2026-08-13 18:39:29 +08:00
2026-08-18 22:47:31 +08:00
2026-08-12 13:03:45 +08:00
2026-08-18 17:17:03 +08:00
2026-08-19 21:45:45 +08:00
2026-08-17 19:04:04 +08:00
2026-08-20 10:15:15 +08:00
2026-08-19 18:43:15 +08:00
2026-08-19 16:10:30 +08:00
2026-08-19 15:02:46 +08:00
2026-07-27 10:20:16 +08:00
2026-08-18 22:47:31 +08:00
2026-08-14 17:43:55 +08:00
2026-08-18 22:47:31 +08:00
2026-08-19 21:23:51 +08:00
2026-08-11 19:19:29 +08:00
2026-08-19 19:23:26 +08:00
2026-08-03 19:03:08 +08:00
2026-08-19 15:04:33 +08:00
2026-08-05 22:05:07 +08:00