e0b97363c9
smarter: factual-model routing (beh-10), fetch_url main-content extraction + 8k cap (url-08), get_weather via met.no (spec-016)
2026-07-17 19:15:18 +02:00
5a3623f813
strengthen picture_edit hint: model must use edit for previously shared images
2026-07-13 16:45:15 +02:00
e21c262299
image input pipeline: content-hash cache, data-url vision, picture_edit real
2026-07-13 16:33:48 +02:00
7e6eae10ee
gpt-image-2 multi-image + fix: tools need reasoning_effort none on gpt-5.6 (ggg mute bug)
2026-07-13 16:17:00 +02:00
ae870db181
human behavior: classifier gate, pacing, splitting, quiet hours, stable prompt prefix
2026-07-13 15:55:26 +02:00
3d2289496b
structured memory: facts/pinned/episodes, batched consolidation, participant-scoped recall
2026-07-13 14:12:21 +02:00
f6c3e7d8e5
pickle -> sqlite store: D6 async writes, D9 reload race, one-shot pickle migration
2026-07-13 13:04:32 +02:00
6e5abf3d2c
fix defects D1-D12 batch 1, structured envelope, saf gates + ops kill-switches
2026-07-13 12:56:32 +02:00
ok
fbec05dfe9
Fix hanging test and establish comprehensive development environment
...
- Fix infinite retry loop in ai_responder.py that caused test_fix1 to hang
- Add missing picture_edit parameter to all AIResponse constructor calls
- Set up complete development toolchain with Black, isort, Bandit, and MyPy
- Create comprehensive Makefile for development workflows
- Add pre-commit hooks with formatting, linting, security, and type checking
- Update test mocking to provide contextual responses for different scenarios
- Configure all tools for 140 character line length and strict type checking
- Add DEVELOPMENT.md with setup instructions and workflow documentation
🤖 Generated with [Claude Code](https://claude.ai/code )
Co-Authored-By: Claude <noreply@anthropic.com >
2025-08-08 19:07:14 +02:00
ok
0bf1baf341
Fix image handling and do not react to bots or unchanged messages.
2024-03-17 15:10:35 +01:00
ok
29b0d6efca
Add reaction on message edit.
2024-03-17 13:46:37 +01:00
ok
1bff1c9719
Fix reaction handling.
2024-03-17 13:25:02 +01:00
ok
73d9b9184d
Add support for reactions with emojis.
2024-03-17 12:39:15 +01:00
ok
3bdf9d03c6
Change the way, memory is logged.
2024-03-17 12:08:00 +01:00
ok
53ed068519
Add support of reply and do not get borred if prev message is borred. Manage memory.
2024-03-16 23:55:29 +01:00
ok
0c23da7c9e
Changes.
2024-03-16 13:44:46 +01:00
ok
39d7c48ee3
Remove some debug output.
2023-12-02 22:36:48 +01:00
ok
c010603178
Fix and remove some tests, make openai calls cachable in a file.
2023-12-02 22:36:00 +01:00
7bcadecb17
Solve new API thing differently.
2023-11-15 10:23:38 +01:00
a2c7aec1e3
Merge branch 'master' of stage.fjerkroa.no:Fjerkroa/discord_bot
2023-11-14 10:27:29 +01:00
488a8ef174
Support new openai API.
2023-11-14 10:27:03 +01:00
ok
924daf134f
Improve exponantial back off on leanardo drawing a bit, and log the sleeps.
2023-08-22 11:30:30 +02:00
ok
95bc6ce041
Add possibility to set retry-model, if AI request failes because of limits. Improve error handling in leanardo AI drawing.
2023-08-22 11:05:17 +02:00
ok
17386af950
Fix channel names in answers on history.
2023-08-21 16:18:37 +02:00
ok
8c718ad272
Do not historise the boreness question.
2023-08-21 13:38:43 +02:00
ok
9d1d99ac50
Try to improve the auto translation.
2023-08-21 13:28:39 +02:00
8a5cef4deb
Add log output, to see which drawing enginge was used.
2023-08-19 21:23:41 +02:00
ok
4ec20f1b26
Add possibility to use Leonardo.ai for image generation and improve it a bit.
2023-08-19 16:04:21 +02:00
ok
e237800348
Support Leonardo.ai to generate images.
2023-08-19 14:33:04 +02:00
ok
6d9fa0e718
Added exponential backoff on ratelimit error, also print the token usage in output.
2023-07-11 14:47:02 +02:00
ok
112f03a47a
Refactor and improve the code a bit
2023-04-13 18:57:37 +02:00
ok
2508a12b44
Fix history limit handling.
2023-04-13 18:36:06 +02:00
ok
bcfe0e968f
Fix history reducing alghorithm
2023-04-13 12:58:55 +02:00
ok
2ca528e211
Comment the code a bit
2023-04-12 22:32:24 +02:00
ok
7c029038f6
Output the returned answer from ai to log
2023-04-12 20:18:14 +02:00
ok
536930d426
Fix shrink history for short path case.
2023-04-12 19:59:27 +02:00
ok
203ebe72c1
Accept JSON structure as channel name
2023-04-12 19:22:45 +02:00
ok
b26b98611a
Make the JSON porse a bit more robust.
2023-04-12 19:19:20 +02:00
ok
b1ece64874
Implement possibility to answer to a different channel
2023-04-12 17:56:31 +02:00
ok
defe598651
Improve handling of wrong keys in response
2023-04-12 12:47:53 +02:00
ok
77fe7a69d6
Make history messages per channel configurable
2023-04-12 12:23:41 +02:00
ok
2db983c462
Improve history handling
...
- Try to keep at least 3 messages from each channel in the history
- Use post processed messages for the history, instead of the raw
messages from the openai API
2023-04-12 12:19:07 +02:00
ok
79fc0d6cba
Do not return "'none'" from parse_maybe_json.
2023-04-10 15:43:46 +02:00
ok
d29595136f
Output the final answer in the log
2023-04-10 14:41:57 +02:00
ok
961e70f17e
See quotet 'none' and 'null' also as None answer.
2023-04-10 14:34:38 +02:00
ok
9a411a3fed
Handle messages as direct, when bot is mentioned.
2023-04-10 13:50:35 +02:00
ok
24ae6bcf32
Convert markdown links to just links.
2023-03-30 17:33:58 +02:00
ok
9eda81d7eb
Improve parse_maybe_json.
2023-03-30 15:52:11 +02:00
ok
585ecc17c6
Parse response values, when they are JSON.
2023-03-30 15:38:42 +02:00
ok
93c5758e02
Some more fixes, reduce retries.
2023-03-29 19:53:45 +02:00
Oleksandr Kozachuk
ok
Fjerkroa Auto