Commit graph

565 commits

Author SHA1 Message Date
Simon Willison
dff53a9cae Better --help for llm keys get, refs #623 2024-11-11 09:53:24 -08:00
Simon Willison
561784df6e llm keys get command, refs #623 2024-11-11 09:47:13 -08:00
Simon Willison
5d1d723d4b Special case treat audio/wave as audio/wav, closes #603 2024-11-07 17:13:54 -08:00
Simon Willison
febbc04fb6
Run cog -r in PRs, use that to update logging.md with new tables (#616)
* Create cog.yml
* Document attachments and prompt_attachments table schemas

Closes #615

---------

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
2024-11-06 06:56:19 -08:00
Simon Willison
98d2c19876 Promote alternative model providers in llm --help 2024-11-06 06:38:53 -08:00
Simon Willison
3352eb9f57 Serialize usage to JSON properly, closes #614 2024-11-06 03:27:25 -08:00
Simon Willison
245e025270 Ran cog, refs #612 2024-11-05 23:45:17 -08:00
Chris Mungall
3b2e5263a3
Allow passing of can_stream in openai_models.py (#600)
* Allow passing of can_stream in openai_models.py

Fixes #599 

* Only set can_stream: false if it is false

Refs https://github.com/simonw/llm/pull/600#issuecomment-2458825866

* Docs for can_stream: false

---------

Co-authored-by: Simon Willison <swillison@gmail.com>
2024-11-05 23:04:13 -08:00
Simon Willison
7146fe82d1 Back to the deprecated Pydantic thing to get tests passing
Refs #612
2024-11-05 23:03:31 -08:00
Simon Willison
12df1a3b2a Show attachment types in llm models --options, closes #612 2024-11-05 22:49:26 -08:00
Simon Willison
1a60fa1667 Test to exercise gpt-4o-audio-preview, closes #608 2024-11-05 21:50:00 -08:00
Simon Willison
0cc4072bcd Support attachments without prompts, closes #611 2024-11-05 21:27:18 -08:00
Simon Willison
41cb5c3387 Ran cog, refs #608 2024-11-05 21:13:36 -08:00
Simon Willison
336ab1013d gpt-4o-audio-preview audio input, refs #608 2024-11-05 21:08:52 -08:00
Simon Willison
fe1e09706f
llm-lambda-labs
!stable-docs
2024-11-04 10:26:02 -08:00
Simon Willison
a7eedd1d2e mypy fix, refs #601 2024-11-01 14:21:01 -07:00
Simon Willison
d7b395bb5b Release 0.17.1 2024-11-01 14:19:39 -07:00
Simon Willison
122265a3d2 Fix for chat continuation in llm chat, refs #601 2024-11-01 14:17:25 -07:00
Simon Willison
a44ba49c21 Release 0.17
Refs #587, #590, #591
2024-10-28 19:36:12 -07:00
Simon Willison
39d61d433a Automated tests for attachments, refs #587 2024-10-28 19:21:11 -07:00
Simon Willison
389acdf52c Track usage on OpenAI stream requests, closes #591 2024-10-28 17:40:40 -07:00
Simon Willison
ba1ccb3a4a Release 0.17a0
Refs #587, #590
2024-10-28 15:46:52 -07:00
Simon Willison
758ff9ac17 Upgrade to pytest-httpx>=0.33.0 2024-10-28 15:41:34 -07:00
Simon Willison
be2953e6ab Ruff and mypy fixes 2024-10-28 15:41:34 -07:00
Simon Willison
db1d77f486 assert_all_responses_were_requested=False 2024-10-28 15:41:34 -07:00
Simon Willison
1f822d820b Update docs with cog 2024-10-28 15:41:34 -07:00
Simon Willison
f0ed54abf1 Docs for CLI attachments, refs #587 2024-10-28 15:41:34 -07:00
Simon Willison
570a3eccae Python attachment documentation, plus fixed a mimetype detection bug
Refs #587
2024-10-28 15:41:34 -07:00
Simon Willison
286cf9fcd9 attachments= keyword argument, tests pass again - refs #587 2024-10-28 15:41:34 -07:00
Simon Willison
a9bc1c7329 llm logs --json for attachments, refs #587 2024-10-28 15:41:34 -07:00
Simon Willison
2384fd52a0 Tighter log output for binary content 2024-10-28 15:41:34 -07:00
Simon Willison
bb5b802d4f llm logs markdown support for attachments, refs #587 2024-10-28 15:41:34 -07:00
Simon Willison
1126393ba1 Docs for writing models that accept attachments, refs #587 2024-10-28 15:41:34 -07:00
Simon Willison
a68af9c8e6 Don't redact non-data URLs for OpenAI models
Refs #587
2024-10-28 15:41:34 -07:00
Simon Willison
cd722f653b Redact base64 data from _prompt_json, refs #587 2024-10-28 15:41:34 -07:00
Simon Willison
dff5b456fd Got llm --continue to work with images, refs #587 2024-10-28 15:41:34 -07:00
Simon Willison
c0fe719df6 Store prompt attachments in attachments and prompt_attachments tables
Refs https://github.com/simonw/llm/issues/587#issuecomment-2439791231
2024-10-28 15:41:34 -07:00
Simon Willison
6df00f92ff First working prototype of new attachments feature, refs #587 2024-10-28 15:41:34 -07:00
Andrew Wason
a466ddf3cd
Fix broken tests. Drop python 3.8. (#580)
* Fix broken tests.  Drop python 3.8.
* Test on Python 3.13 too

---------

Co-authored-by: Simon Willison <swillison@gmail.com>
2024-10-27 11:26:47 -07:00
Simon Willison
7e6031e382
llm-gguf, llm-jq
!stable-docs
2024-10-26 22:44:06 -07:00
Simon Willison
d654c95212 Release notes for 0.16 2024-09-12 16:20:12 -07:00
Simon Willison
38239839ae
Release 0.16
Refs #552, #553, #554, #558, #567, #570, #573
2024-09-12 16:09:46 -07:00
Simon Willison
bfcfd2c91b
o1-preview and o1-mini, refs #570 (#573) 2024-09-12 16:08:04 -07:00
Kian-Meng Ang
50520c7c1c
Fix typos (#567)
Found via `codespell -H -L wit,thre`

!stable-docs
2024-09-08 08:44:43 -07:00
Simon Willison
7d6ece2a31 Fix for broken markdown on openai-models page
Refs #558 !stable-docs
2024-08-25 18:03:46 -07:00
Simon Willison
e867e13d1b
Use model_dump() instead of dict()
Closes #554
2024-08-18 17:39:09 -07:00
Simon Willison
6deed8f976 get_model() improvement, get_default_model() / set_default_wodel() now documented
Refs #553
2024-08-18 17:37:31 -07:00
Simon Willison
24cc042e59 Use self.get_key() in OpenAI Chat class, refs #552 2024-08-18 17:01:28 -07:00
Simon Willison
d075336c69 Release 0.15
Refs #515, #525, #536, #537
2024-07-18 12:31:14 -07:00
Simon Willison
562fefb374 Use 3-small in docs instead of ada-002
Spotted while working on #537
2024-07-18 12:23:49 -07:00