Commit Graph

48 Commits

Author SHA1 Message Date
LeonOstrez
4c1fe834a8 fix getting args.root, ask feedback only on success and keyboard interrupt (not on exceptions), fix terminal width and erasing old logs 2023-09-29 13:31:11 +01:00
Nicholas Albion
7bf5de8861 Improved logging 2023-09-29 18:18:55 +10:00
Nicholas Albion
2aaa914c26 Merge branch 'main' into debugging_ipc
# Conflicts:
#	README.md
#	pilot/helpers/agents/CodeMonkey.py
#	pilot/helpers/agents/Developer.py
#	pilot/prompts/system_messages/architect.prompt
#	pilot/utils/arguments.py
#	pilot/utils/llm_connection.py
#	pilot/utils/utils.py
2023-09-28 13:53:49 +10:00
Nicholas Albion
108271f28a ignore ': OPENROUTER PROCESSING' 2023-09-27 13:10:52 +10:00
Nicholas Albion
6dd5a032fa Improved JSON prompting for GPT-4 and recover incomplete JSON responses from Code Llama 2023-09-27 10:53:44 +10:00
Nicholas Albion
e5e0f56d2e JSON validation working 2023-09-26 19:38:24 +10:00
Nicholas Albion
8a024c2ff2 rejecting responses that are not JSON.
Need to fix prompts for GPT-4
2023-09-26 17:27:54 +10:00
Nicholas Albion
b317f58550 clean up 2023-09-23 16:27:02 +10:00
Nicholas Albion
156b361263 function_call-style JSON response from gpt-4, gpt-3_5, codellama, palm-2-chat-bison 2023-09-23 00:45:23 +10:00
Nicholas Albion
d4879a04b7 ARCHITECTURE function_calls works on meta-llama/codellama-34b-instruct 2023-09-21 23:05:37 +10:00
Nicholas Albion
2b4186a4f8 test_chat_completion_Architect passes 2023-09-20 23:15:37 +10:00
Nicholas Albion
4b7aa2df22 get_prompt() moved from llm_connection to utils and works from unit tests 2023-09-20 22:17:37 +10:00
Nicholas Albion
0660244017 fabulous does not work on Windows 2023-09-20 22:14:59 +10:00
Zvonimir Sabljic
ffe4fbeba9 Enabled catching of max token limit errors from OpenAI's response 2023-09-18 19:18:54 -07:00
LeonOstrez
67209b5b27 merge master into debugging_ipc branch 2023-09-18 19:09:37 -07:00
Zvonimir Sabljic
65135344ab Converted colored leftovers 2023-09-15 09:43:30 +02:00
Zvonimir Sabljic
9a7c15e0c5 Merge branch 'ipc' into debugging_ipc 2023-09-14 09:40:36 +02:00
Zvonimir Sabljic
151aa051e2 Improved debugging process and enabled splitting of app development into tasks and then into steps
- split step implementation into different functions
- standardized the return value in the implementation process - { "success": True }
- added propagation of errors back to the recursion level 0 with TooDeepRecursionError and TokenLimitError
- created new class Debugger and moved debugging in it
2023-09-12 21:32:56 +02:00
Zvonimir Sabljic
19ac692509 Don't send max_tokens to openai api so we can use as much context as possible 2023-09-12 21:28:01 +02:00
Nicholas Albion
54bee95a8c Merge remote-tracking branch 'origin/main'
# Conflicts:
#	README.md
#	pilot/helpers/AgentConvo.py
#	pilot/utils/llm_connection.py
2023-09-12 22:57:48 +10:00
LeonOstrez
b023205a53 Merge pull request #85 from alter123/patch-1
Add check when the response is empty
2023-09-12 08:40:07 +02:00
LeonOstrez
9ffcdf79bb Merge pull request #82 from nalbion/feature/test_CodeMonkey
Feature/test code monkey
2023-09-12 08:09:12 +02:00
Jay
74cbe33421 Add check when the response is empty
{
  "id": "",
  "object": "",
  "created": 0,
  "model": "",
  "prompt_annotations": [
    {
      "prompt_index": 0,
      "content_filter_results": {
        "hate": { "filtered": false, "severity": "safe" },
        "self_harm": { "filtered": false, "severity": "safe" },
        "sexual": { "filtered": false, "severity": "safe" },
        "violence": { "filtered": false, "severity": "safe" }
      }
    }
  ],
  "choices": [],
  "usage": null
}


In newer versions of models, at times choices array can be empty
2023-09-12 00:01:13 +05:30
Nicholas Albion
a94cbf9209 added documentation 2023-09-11 22:15:26 +10:00
Nicholas Albion
a8008b263e Merge branch 'feature/prompts_and_agents_documentation' 2023-09-11 14:18:10 +10:00
Nicholas Albion
578a665f19 Merge remote-tracking branch 'origin/main'
# Conflicts:
#	pilot/helpers/agents/Developer.py
#	pilot/utils/arguments.py
#	pilot/utils/llm_connection.py
2023-09-11 14:15:11 +10:00
Goon
3b207987aa Merge remote-tracking branch 'upstream/main' 2023-09-11 10:26:05 +07:00
Goon
c39346868a fix(gitignore): rm cache 2023-09-11 10:08:57 +07:00
Goon
367caa1797 fix(gitignore): rm pilot-env and cache 2023-09-11 10:08:42 +07:00
Nicholas Albion
da1e0feb3b added documentation 2023-09-11 03:05:39 +10:00
Nicholas Albion
a38c7c4f6d linting 2023-09-09 11:54:16 +10:00
Nicholas Albion
4b64631bec linting 2023-09-09 11:50:55 +10:00
Nicholas Albion
425d6ac0bf spinner for LLM timeout 2023-09-09 10:43:59 +10:00
Zvonimir Sabljic
6a46851b20 Added catching and retrying when there is an error with Token limit 2023-09-08 18:03:54 +02:00
Goon
02623b6353 fix(llm connection): add openrouter api endpoint 2023-09-08 15:52:29 +07:00
Nicholas Albion
48edfae03c handle rate_limit_exceeded error 2023-09-08 03:05:26 +10:00
LeonOstrez
83378033be remove checking for tokens before openai api request and handle too many tokens in response 2023-09-06 16:46:08 +02:00
Zvonimir Sabljic
83ebd7939d Merge branch 'main' into sander110419-main 2023-09-05 22:52:01 +02:00
Zvonimir Sabljic
a9ead6ecbb Fix to enable regular OpenAI access 2023-09-05 22:50:48 +02:00
Dani Acosta
af6a972cba Add OPENAI_MODEL env var
Adds a env variable OPENAI_MODEL to be able to use different models to GPT-4
2023-09-05 00:15:21 +02:00
Zvonimir Sabljic
28d0143536 TEMP fix 2023-09-01 18:29:31 +02:00
Zvonimir Sabljic
85ac7e8276 Refactored all prints to be colored with fabulous and not termcolor 2023-09-01 18:28:20 +02:00
Zvonimir Sabljic
ca58c4958d Implemented final version of IPC communication 2023-09-01 18:27:00 +02:00
Sander Hilven
a4d520763f Added model selection to .env and update readme 2023-09-01 10:34:12 +02:00
Sander Hilven
660047a071 Hardcoded model in endpoint URL, now fixed. 2023-09-01 09:56:39 +02:00
Sander Hilven
984379fe71 Adde Azure OpenAI endpoint.
Tested and confirmed working.
2023-09-01 09:53:17 +02:00
Zvonimir Sabljic
f4dc07407e Made logging look nicer 2023-08-25 14:21:41 +02:00
Zvonimir Sabljic
8c58d2803d Renamed the repo 2023-08-16 14:57:55 +02:00