RA.Aid/ra_aid/tools
Ariel Frischer 686ab42f88
Add Deepseek Provider Support and Custom Deepseek Reasoner Chat Model (#50)
* chore: Add DeepSeek provider environment variable support in env.py

* feat: Add DeepSeek provider validation strategy in provider_strategy.py

* feat: Add support for DEEPSEEK provider in initialize_llm function

* feat: Create ChatDeepseekReasoner for custom handling of R1 models

* feat: Configure custom OpenAI client for DeepSeek API integration

* chore: Remove unused json import from deepseek_chat.py

* refactor: Simplify invocation_params and update acompletion_with_retry method

* feat: Override _generate to ensure message alternation in DeepseekReasoner

* feat: Add support for ChatDeepseekReasoner in LLM initialization

* feat: Use custom ChatDeepseekReasoner for DeepSeek models in OpenRouter

* fix: Remove redundant condition for DeepSeek model initialization

* feat: Add DeepSeek support for expert model initialization in llm.py

* feat: Add DeepSeek model handling for OpenRouter in expert LLM initialization

* fix: Update model name checks for DeepSeek and OpenRouter providers

* refactor: Extract common logic for LLM initialization into reusable methods

* test: Add unit tests for DeepSeek and OpenRouter functionality

* test: Refactor tests to match updated LLM initialization and helpers

* fix: Import missing helper functions to resolve NameError in tests

* fix: Resolve NameError and improve environment variable fallback logic

* feat(readme): add DeepSeek API key requirements to documentation for better clarity on environment variables
feat(main.py): include DeepSeek as a supported provider in argument parsing for enhanced functionality
feat(deepseek_chat.py): implement ChatDeepseekReasoner class for handling DeepSeek reasoning models
feat(llm.py): add DeepSeek client creation logic to support DeepSeek models in the application
feat(models_tokens.py): define token limits for DeepSeek models to manage resource allocation
fix(provider_strategy.py): correct validation logic for DeepSeek environment variables to ensure proper configuration
chore(memory.py): refactor global memory structure for better readability and maintainability in the codebase
2025-01-22 07:21:10 -05:00
..
__init__.py Add Tavily tool. 2024-12-23 16:48:47 -05:00
agent.py Increase recusion limit; Handle 429 better; Improve prompts. 2025-01-09 15:52:51 -05:00
expert.py Adding Gemini API due to openrouter's limitations. (#34) 2025-01-04 07:19:05 -05:00
file_str_replace.py improve file_str_replace status reporting 2024-12-17 07:46:16 -05:00
fuzzy_find.py improve programmer prompt 2024-12-18 12:29:01 -05:00
human.py Version bump 2024-12-23 15:53:23 -05:00
list_directory.py Changes (#37) 2025-01-09 09:46:32 -05:00
memory.py Add Deepseek Provider Support and Custom Deepseek Reasoner Chat Model (#50) 2025-01-22 07:21:10 -05:00
programmer.py Add Aider config File Argument Support (#43) 2025-01-21 13:17:59 -05:00
read_file.py Make read_file always output status panel. 2024-12-27 19:55:15 -05:00
reflection.py Extract tool reflection. 2025-01-09 15:35:14 -05:00
research.py Extricate research subtasks. 2024-12-21 11:27:14 -05:00
ripgrep.py Changes (#37) 2025-01-09 09:46:32 -05:00
shell.py allow enabling of cowboy mode from approval prompt; print tool error panels 2024-12-17 15:06:51 -05:00
web_search_tavily.py Tavily integration 2024-12-23 17:20:24 -05:00
write_file.py make write_file tool available and fix empty dir creation bug 2024-12-17 08:13:04 -05:00