yujunjun/LightRAG
1
0
Fork 0
mirror of https://github.com/HKUDS/LightRAG.git synced 2025-11-30 08:56:23 +00:00
Code Issues Packages Projects Releases Wiki Activity
4,749 Commits 7 Branches 72 Tags
Commit Graph

145 Commits

Author SHA1 Message Date
yangdx
0be4f0144b Merge branch 'entityTypesServerSupport' 2025-08-27 12:23:58 +08:00
yangdx
ff0a18e08c Unify SUMMARY_LANGUANGE and ENTITY_TYPES implementation method 2025-08-27 12:23:22 +08:00
yangdx
cb0a035076 Update env.example 2025-08-27 11:12:52 +08:00
Thibo Rosemplatt
c3aabfc251 Merge branch 'main' into entityTypesServerSupport 2025-08-26 21:48:20 +02:00
yangdx
6bcfe696ee feat: add output length recommendation and description type to LLM summary 
- Add SUMMARY_LENGTH_RECOMMENDED parameter (600 tokens)
- Optimize prompt temple for LLM summary
 2025-08-26 14:41:12 +08:00
yangdx
84416d104d Increase default LLM summary merge threshold from 4 to 8 for reducing summary trigger frequency 2025-08-26 03:57:35 +08:00
yangdx
de2daf6565 refac: Rename summary_max_tokens to summary_context_size, comprehensive parameter validation for summary configuration 
- Update algorithm logic in operate.py for better token management
- Fix health endpoint to use correct parameter names
 2025-08-26 01:35:50 +08:00
Thibo Rosemplatt
d054ec5d00 Added entity_types as a user defined variable (via .env) 2025-08-23 20:16:11 +02:00
yangdx
3d5e6226a9 Refactored rerank_example file to utilize the updated rerank function. 2025-08-23 22:51:41 +08:00
yangdx
9bc349ddd6 Improve Empty Keyword Handling logic 2025-08-23 11:50:58 +08:00
yangdx
1be9a54c8d Rename ENABLE_RERANK to RERANK_BY_DEFAULT and update default to true 2025-08-23 09:46:51 +08:00
yangdx
47485b130d refac(ui): Show rerank binding info on status card 
- Remove separate ENABLE_RERANK flag in favor of rerank_binding="null"
- Change default rerank binding from "cohere" to "null" (disabled)
- Update UI to display both rerank binding and model information
 2025-08-23 02:04:14 +08:00
yangdx
580cb7906c feat: Add multiple rerank provider support to LightRAG Server by adding new env vars and cli params 
- Add --enable-rerank CLI argument and ENABLE_RERANK env var
- Simplify rerank configuration logic to only check enable flag and binding
- Update health endpoint to show enable_rerank and rerank_configured status
- Improve logging messages for rerank enable/disable states
- Maintain backward compatibility with default value True
 2025-08-22 19:29:45 +08:00
yangdx
16a1ef1178 Update summary_max_tokens default from 10k to 30k tokens 2025-08-21 23:16:07 +08:00
yangdx
718025dbea Update embedding configuration docs and add aws_bedrock option 2025-08-21 17:55:04 +08:00
yangdx
4b2ef71c25 feat: Add extra_body parameter support for OpenRouter/vLLM compatibility 
- Enhanced add_args function to handle dict types with JSON parsing
- Added reasoning and extra_body parameters for OpenRouter/vLLM compatibility
- Updated env.example with OpenRouter/vLLM parameter examples
 2025-08-21 13:06:28 +08:00
yangdx
5d34007f2c Add presence penalty config option for smaller models 
- Add OPENAI_LLM_PRESENCE_PENALTY setting
- Recommend 1.5 for Qwen3 <32B params
- Update max completion tokens comment
 2025-08-21 11:35:23 +08:00
yangdx
0dd245e847 Add OpenAI reasoning effort and max completion tokens config options 2025-08-21 11:04:06 +08:00
yangdx
0e67ead8fa Rename MAX_TOKENS to SUMMARY_MAX_TOKENS for clarity 2025-08-21 10:15:20 +08:00
yangdx
aa22772721 Refactor LLM temperature handling to be provider-specific 
• Remove global temperature parameter
• Add provider-specific temp configs
• Update env example with new settings
• Fix Bedrock temperature handling
• Clean up splash screen display
 2025-08-20 23:52:33 +08:00
yangdx
df7bcb1e3d Add LLM_TIMEOUT configuration for all LLM providers 
- Add LLM_TIMEOUT env variable
- Apply timeout to all LLM bindings
 2025-08-20 23:50:57 +08:00
yangdx
4c556d8aae Set default TIMEOUT value to 150, and gunicorn timeout to TIMEOUT+30 2025-08-20 22:04:32 +08:00
yangdx
d5e8f1e860 Update default query parameters for better performance 
- Increase chunk_top_k from 10 to 20
- Reduce max_entity_tokens to 6000
- Reduce max_relation_tokens to 8000
- Update web UI default values
- Fix max_total_tokens to 30000
 2025-08-18 19:32:11 +08:00
yangdx
da7e4b79e5 Update documentation in README files 2025-08-17 02:23:14 +08:00
yangdx
2a781dfb91 Update Neo4j database naming in env.example 2025-08-15 19:14:38 +08:00
yangdx
6cab68bb47 Improve KG chunk selection documentation and configuration clarity 2025-08-15 10:09:44 +08:00
yangdx
2a46667ac9 Add OpenAI frequency penalty sample env params 2025-08-14 01:50:27 +08:00
yangdx
5a40ff654e Change KG chunk selection default to VECTOR 
- Set KG_CHUNK_PICK_METHOD default to VECTOR
- Update env.example with new config option
 2025-08-13 23:10:42 +08:00
yangdx
095e0cbfa2 Refac: Add workspace infomation to all logger output for all storage type 2025-08-12 01:19:09 +08:00
yangdx
f5ac6a9f4b Add default Ollama embedding context length 
- Set default context length to 8192
- Overide the default context lenght for LLM in binding_options.py
 2025-08-08 13:51:25 +08:00
yangdx
dec4148075 Merge branch 'main' into Matt23-star/main 2025-08-08 02:24:34 +08:00
yangdx
f38e10559e Update PostgreSQL vector index configuration 
- Remove FLAT index support
- Standardize on HNSW as default
- Add dimension validation
- Improve error logging
- Clean up index creation code
 2025-08-08 02:21:06 +08:00
yangdx
f4ef254de2 fix(neo4j): enhance connection lifecycle management to prevent timeout errors 
- Add max_connection_lifetime, liveness_check_timeout, keep_alive parameters
- Extend retry mechanisms for connection reset scenarios
- Update config examples with new Neo4j connection options
- Resolves ClientTimeoutException during data insertion operations
 2025-08-08 01:07:45 +08:00
Matt23-star
727ca43d3c feat: add vector index creation functionality for PostgreSQL 2025-08-07 23:07:18 +08:00
yangdx
7780776af6 Update env.example 2025-08-06 18:50:58 +08:00
yangdx
8dd3069005 feat: add tiktoken cache directory support for offline deployment 
- Mount tiktoken cache volume in docker-compose
- Add TIKTOKEN_CACHE_DIR environment variable
- Update env.example with tiktoken cache configuration
- Added /data to gitignore
- Added tiktoken env example
 2025-08-05 14:22:19 +08:00
yangdx
c5babf61d7 Feat: Change embedding formats from float to base64 for efficiency 
- Add base64 support for Jina embeddings
- Add base64 support for OpenAI embeddings
- Update env.example with new embedding options
 2025-08-05 11:38:40 +08:00
yangdx
adf7ec8e35 feat: Add OpenAI LLM Options support with BindingOptions framework 
- Add OpenAILLMOptions dataclass with full OpenAI API parameter support
- Integrate OpenAI options in config.py for automatic binding detection
- Update server functions to inject OpenAI options for openai/azure_openai bindings
 2025-08-05 03:47:26 +08:00
yangdx
3099748668 Add temperature fallback for Ollama LLM binding 
- Implement OLLAMA_LLM_TEMPERATURE env var
- Fallback to global TEMPERATURE if unset
- Remove redundant OllamaLLMOptions logic
- Update env.example with new setting
 2025-08-05 01:50:09 +08:00
yangdx
e5e3f0f878 Fix(Ollama option): change stop option from string to list and add fallback global temperature setting 2025-08-04 19:43:14 +08:00
yangdx
9d5603d35e Set the default LLM temperature to 1.0 and centralize constant management 2025-07-31 17:15:10 +08:00
yangdx
3c530b21b6 Update README 2025-07-31 13:00:09 +08:00
yangdx
c6bd9f0329 Disable conversation history by default 
- Set default history_turns to 0
- Mark history_turns as deprecated
- Remove history_turns from example
- Update documentation comments
 2025-07-31 12:28:42 +08:00
yangdx
9923821d75 refactor: Remove deprecated max_token_size from embedding configuration 
This parameter is no longer used. Its removal simplifies the API and clarifies that token length management is handled by upstream text chunking logic rather than the embedding wrapper.
 2025-07-29 10:49:35 +08:00
yangdx
75d1b1e9f8 Update Ollama context length configuration 
- Rename OLLAMA_NUM_CTX to OLLAMA_LLM_NUM_CTX
- Increase default context window size
- Add requirement for minimum context size
- Update documentation examples
 2025-07-29 09:53:37 +08:00
yangdx
f2ffff063b feat: refactor ollama server configuration management 
- Add ollama_server_infos attribute to LightRAG class with default initialization
- Move default values to constants.py for centralized configuration
- Refactor OllamaServerInfos class with property accessors and CLI support
- Update OllamaAPI to get configuration through rag object instead of direct import
- Add command line arguments for simulated model name and tag
- Fix type imports to avoid circular dependencies
 2025-07-28 01:38:35 +08:00
yangdx
519e81aaeb Update env.example 2025-07-27 18:21:10 +08:00
yangdx
ebaff228aa feat: Add rerank score filtering with configurable threshold 
- Add DEFAULT_MIN_RERANK_SCORE constant (default: 0.0)
- Add MIN_RERANK_SCORE environment variable support
- Filter chunks with rerank scores below threshold in process_chunks_unified
- Add info-level logging for filtering operations
- Handle empty results gracefully after filtering
- Maintain backward compatibility with non-reranked chunks
 2025-07-27 16:37:44 +08:00
yangdx
055629d30d Reduce default max total tokens to 30k 2025-07-27 10:33:06 +08:00
yangdx
b3c2987006 Reduce default MAX_TOKENS from 32000 to 10000 2025-07-26 08:13:49 +08:00