Commit Graph
86 Commits
Author SHA1 Message Date
Tasha Upchurch a79cc6de38 Handle running event loop in LiteLLM prompt handler 2025-09-22 20:20:05 -06:00
Tasha Upchurch 0a5e44cafc Merge pull request #5 from Hopping-Mad-Games/codex/update-litellmmodelprovideradv-to-fetch-current-models-md1tqt
Handle OpenAI responses models requiring max_completion_tokens
2025-09-22 17:37:20 -06:00
Tasha Upchurch e990930363 Merge branch 'main' into codex/update-litellmmodelprovideradv-to-fetch-current-models-md1tqt 2025-09-22 17:35:52 -06:00
Tasha Upchurch 0807ebbad8 Handle OpenAI responses models requiring max_completion_tokens 2025-09-22 17:33:46 -06:00
Tasha Upchurch 41f97a986b Merge pull request #4 from Hopping-Mad-Games/codex/update-litellmmodelprovideradv-to-fetch-current-models
Improve OpenAI model discovery
2025-09-22 12:02:30 -06:00
Tasha Upchurch 23583f0eb7 Improve OpenAI model discovery 2025-09-22 12:02:07 -06:00
TashaSkyUp 7cf82171e5 Make environment management non-polluting and safe
This commit implements a clean architecture where environment variables
are only read, never set, ensuring compatibility with other ComfyUI
packages.

The changes: - Remove all code that sets global environment variables -
Clean up config.yaml to remove API key placeholders - Add documentation
for safe environment handling - Add comprehensive tests to verify
non-polluting behavior - Ensure backward compatibility for other
packages

The commit includes detailed docs explaining the clean architecture and
safety guarantees for package compatibility.
2025-06-25 17:22:53 -06:00
TashaSkyUp fa7789a035 Refactor config handling and environment variable setup
- Add get_api_key and get_config_value helpers - Only set internal
config keys as environment variables with a prefix - Avoid polluting
global environment with API keys - Improve code organization and
comments
2025-06-24 18:53:40 -06:00
TashaSkyUp 40831eaa28 Add agent node docstrings, improve error handling, and update config
- Add detailed docstrings to AgentNode, BasicRecursionFilterNode, and
DocumentChunkRecursionFilterNode for clarity and documentation. -
Improve input validation and error handling in agent nodes. - Update
.gitignore to cover more environment, config, and IDE files. - Refactor
LiteLLMCustomEndpointProvider for custom API endpoints. - Add
conditional imports and dummy functions for LightRAG when unavailable. -
Update config.yaml to use environment variables for API keys and clarify
usage. - Refine requirements.txt to avoid heavy LightRAG dependencies. -
Add tests/__init__.py with test suite metadata; remove redundant test
files. - Fix CustomDict import and model config handling in
litellmnodes.py. - Improve chunk processing and recursion filter logic
in agent nodes. - Update utils/__init__.py to include env_config.
2025-06-24 17:47:36 -06:00
TashaSkyUp b84323f12d update 2025-06-17 17:52:46 -06:00
TashaSkyUp f434f40e9b Enhance LiteLLMCompletionListOfPrompts to support dual processing modes for prompts and message histories 2025-04-03 18:15:20 -06:00
TashaSkyUp 7c3d09189d Refactor messages input definition to simplify optional handling 2025-04-03 17:56:44 -06:00
TashaSkyUp 0663290028 Refactor LiteLLMCompletionListOfPrompts to enhance message handling and input structure 2025-04-03 17:43:12 -06:00
TashaSkyUp 6c5f450b00 Enhance LiteLLMCompletion to support message history input for prompts 2025-04-03 17:30:48 -06:00
TashaSkyUp 127a96a33f Refactor PDF processing to use PyMuPDF for improved performance and error handling; update LiteLLMCompletion to return messages alongside completions. 2025-04-03 16:54:47 -06:00
TashaSkyUp 9f0c0c0a22 possible fix for event loop being closed error 2025-03-02 12:45:22 -07:00
TashaSkyUp feca2a205b fix for typo 2025-02-05 15:38:35 -07:00
TashaSkyUp f3c0c1ee7f fix 2025-02-05 15:24:44 -07:00
TashaSkyUp 21e897645d debugging messages for broken agent node. 2025-02-05 15:15:49 -07:00
TashaSkyUp 43ae1a3104 debugging messages for broken agent node. 2025-02-05 15:11:28 -07:00
TashaSkyUp 4350efc607 debugging messages for broken agent node. 2025-02-05 14:54:44 -07:00
TashaSkyUp 25adf1f474 fix for broken inference on non "o" models 2025-02-05 14:00:56 -07:00
TashaSkyUp a05b3933f6 reasoning_effort for o1 o3 2025-02-04 19:53:16 -07:00
TashaSkyUp c8c8821039 better o1 and o3 support 2025-02-04 18:49:03 -07:00
TashaSkyUp 0239c3917c limited o1 and o3 support 2025-02-04 18:38:53 -07:00
TashaSkyUp 5b75b6155c some debug print statements for lightrag 2025-01-27 14:15:48 -07:00
TashaSkyUp 7434138858 some debug print statements for lightrag 2025-01-27 13:47:36 -07:00
TashaSkyUp 9545fe6624 fix for invalid characters for lightrag_nodes.py 2025-01-13 08:01:21 -07:00
TashaSkyUp 041c9bad54 attempt to prevent some litellm spam in the console 2025-01-09 20:15:25 -07:00
TashaSkyUp a9cf8acec3 image completion provider. 2025-01-09 18:51:24 -07:00
TashaSkyUp 754c86e20b image completion provider. 2025-01-09 18:48:44 -07:00
TashaSkyUp 7f3d33e7cd image completion provider. 2025-01-09 18:41:53 -07:00
TashaSkyUp e81f570819 image completion provider. 2025-01-09 18:07:39 -07:00
TashaSkyUp 9c020366d2 adding support for togetherai: meta-llama/Llama-3.2-90B-Vision-Instruct-Turbo 2025-01-09 17:39:33 -07:00
TashaSkyUp 9cf6490047 adding support for togetherai: meta-llama/Llama-3.2-90B-Vision-Instruct-Turbo 2025-01-09 17:31:47 -07:00
TashaSkyUp 6f4ab84bee adding support for togetherai deepseek-v3 2025-01-09 17:18:01 -07:00
TashaSkyUp 62eba628db fix for async event loop missing 2025-01-06 23:23:13 -07:00
TashaSkyUp a487271576 adding chunk recursion filter 2025-01-05 14:34:22 -07:00
TashaSkyUp 32453b239d added ability to have an inner recursion filter on any basic recursion filter 2024-12-01 15:43:15 -07:00
TashaSkyUp a706ca1185 bugfix 2024-11-30 18:32:40 -07:00
TashaSkyUp 94937dacf6 fix for closed event loop bug 2024-11-29 13:35:53 -07:00
TashaSkyUp a6dace5c82 added missing agents module 2024-11-28 10:39:07 -07:00
TashaSkyUp 2ed779f749 fix for vision in v2 litellm node 2024-11-27 22:05:00 -07:00
TashaSkyUp 1bf860bda2 refactor for consistency with pep 8 naming conventions 2024-11-27 10:25:04 -07:00
TashaSkyUp 5995fdb6eb fixes for caching/saving pydantic models for structured completions. 2024-11-22 11:03:31 -07:00
TashaSkyUp 3031569545 added missing modules 2024-11-22 09:46:09 -07:00
TashaSkyUp 7b9207eed2 feat: Add response caching for AgentNode
- Implement JSON-based response caching system
- Add use_last_response parameter for cache control
- Move cache directory to data folder
- Remove unused safe_print utilities
- Fix messages_results typo
2024-11-22 09:42:11 -07:00
TashaSkyUp 856b64a3be fix for multiple prompts and message history(output) 2024-11-21 19:45:28 -07:00
TashaSkyUp 0f405cba7a fix for maintaining structured outputs 2024-11-21 14:16:25 -07:00
TashaSkyUp 2b921a216e feat: Enhance Agent prompting and memory provider
- Add template for memory prompt generation with focus on retrieving relevant context
- Update BasicRecursionFilter with date-aware prompting and XML tags
- Add LLLM provider support to AgentMemoryProvider for prompt refinement
2024-11-21 08:24:48 -07:00