gmakstutis/LightRAG
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
4804 commits 5 branches 72 tags 705 MiB
Commit graph

3104 commits

Author SHA1 Message Date
yangdx
c86f863fa4 feat: optimize entity extraction for smaller LLMs 
Simplify entity relationship extraction process to improve compatibility
and performance with smaller, less capable language models.

Changes:
- Remove iterative gleaning loop with LLM-based continuation decisions
- Simplify to single gleaning pass when entity_extract_max_gleaning > 0
- Streamline entity extraction prompts with clearer instructions
- Add explicit completion delimiter signals in all examples
 2025-09-03 10:33:01 +08:00
yangdx
9d81cd724a Fix typo: change "Equiment" to "Equipment" in entity types 2025-09-02 03:19:31 +08:00
yangdx
476b64c9d4 Update webui assets 2025-09-02 03:03:19 +08:00
yangdx
4e37ff5f2f Bump API verstion to 0212 2025-09-02 03:02:39 +08:00
yangdx
5b2deccbef Improve text normalization and add entity type capitalization 
- Capitalize entity types with .title()
- Add non-breaking space handling
- Add narrow non-breaking space regex
 2025-09-02 02:51:41 +08:00
yangdx
29f0ecc88c Refactor entity extraction prompts and remove completion delimiter 
• Update prompt structure and wording
• Remove deprecated completion delimiter
• Add quality guidelines section
• Improve instruction clarity
• Enhance continue extraction prompt
 2025-09-02 02:14:14 +08:00
yangdx
3f8a9abe7e Refactor extraction result processing to reduce code duplication 
• Extract shared processing logic
• Add delimiter pattern fixes
• Improve bracket standardization
 2025-09-02 01:22:29 +08:00
yangdx
3cdc98f366 Improve extraction parsing with better bracket handling and delimiter fixes 
• Standardize Chinese/English brackets
• Fix incomplete tuple delimiters
• Remove duplicate delimiter fix code
• Support mixed bracket formats
• Enhance record parsing robustness
 2025-09-02 00:26:04 +08:00
yangdx
8bbf307aeb Fix regex to match multiline content in extraction parsing 
• Remove non-greedy quantifier
• Add DOTALL flag for multiline matching
• Apply to both parsing functions
• Enable cross-line content extraction
 2025-09-01 10:35:06 +08:00
yangdx
7baeb186c6 Fix regex to use non-greedy matching for parentheses extraction 2025-09-01 10:10:45 +08:00
yangdx
692357fbf3 Add conflict resolution instruction to entity summarization prompt 
- Add conflict handling step
- Handle entities with same name
- Separate then consolidate summaries
 2025-09-01 08:51:19 +08:00
yangdx
e95622ca7b fix(utils): enhance remove_think_tags to handle orphaned </think> closing tags 
The function now properly handles cases where text contains </think> closing tags
without corresponding <think> opening tags, which can occur due to content
truncation or processing errors.
 2025-09-01 07:17:30 +08:00
yangdx
30be70991d Bump API version to 0211 2025-09-01 01:23:22 +08:00
yangdx
5fd7682f16 Fix LLM output instability for <|> tuple delimiter 
- Replace <||> with <|>
- Replace < | > with <|>
- Apply fix in both functions
- Handle delimiter variations
- Improve parsing reliability
 2025-09-01 01:22:27 +08:00
yangdx
ec059d1b5d Fix typo and clarify delimiter formatting in relationship extraction Prompt 
- Fix "feild" → "field" typo
- Clarify delimiter spacing rules
 2025-09-01 00:42:59 +08:00
yangdx
c8c59c38b0 Fix entity types configuration to support JSON list parsing 
- Add JSON parsing for list env vars
- Update entity types example format
- Add list type support to get_env_value
 2025-09-01 00:14:57 +08:00
yangdx
1a015a7015 Add queue_name parameter to priority_limit_async_func_call for better logging 
• Add queue_name parameter to decorator
• Update all log messages with queue names
• Pass specific names for LLM and embedding
 2025-08-31 23:47:22 +08:00
yangdx
4e751e0653 refac: Enhance extraction with improved prompts and parser 
-   **Prompts**: Restructured prompts with clearer steps and quality guidelines. Simplified the relationship tuple by removing `relationship_strength`
-   **Model**: Updated default entity types to be more comprehensive and consistently capitalized (e.g., `Location`, `Product`)
 2025-08-31 22:24:11 +08:00
yangdx
75de40da41 Fix typo in relationship extraction log messages 2025-08-31 17:45:16 +08:00
yangdx
97c9600085 Improve extraction error handling and field validation 
• Add field count validation warnings
• Fix relationship field count (5→6)
• Change error logs to warnings
 2025-08-31 17:33:42 +08:00
yangdx
b747417961 feat: enhance text extraction text sanitization and normalization 
- Improve reduntant quotes in entity and relation name, type and keywords
- Add HTML tag cleaning and Chinese symbol conversion
- Filter out short numeric content and malformed text
- Enhance entity type validation with character filtering
 2025-08-31 13:17:20 +08:00
yangdx
d4bbc5dea9 refactor: Merge multi-step text sanitization into single function 2025-08-31 10:36:56 +08:00
yangdx
69890ff2e1 Bump core version to 1.4.8 and api version to 0210 2025-08-31 03:01:33 +08:00
yangdx
8bab240dbc Update webui assets 2025-08-31 03:00:16 +08:00
yangdx
25b5d176cd Fix label selection with leading/trailing whitespace 
• Fix AsyncSelect value trimming issue
• Preserve whitespace in label display
• Use safe keys for command items
• Add GraphControl dependency fix
• Add debug logging for graph labels
 2025-08-31 02:54:39 +08:00
yangdx
ae09b5c656 refactor: eliminate conditional imports and simplify LightRAG initialization 
- Remove conditional import block, replace with lazy loading factory functions
- Add create_llm_model_func() and create_llm_model_kwargs() for clean configuration
- Update wrapper functions with lazy imports for better performance
- Unify LightRAG initialization, eliminating duplicate conditional branches
- Reduce code complexity by 33% while maintaining full backward compatibility
 2025-08-31 00:18:29 +08:00
yangdx
332202c111 Fix lambda closure bug in embedding function configuration 
• Replace lambda with proper async function
• Capture config values at creation time
• Avoid closure variable reference issues
• Add factory function for embeddings
• Remove test file for closure bug
 2025-08-30 23:43:34 +08:00
avchauzov
414d47d12a fix(server): Resolve lambda closure bug in embedding_func 
Fixes #2023. Resolves an issue where the embedding function would incorrectly fall back to the OpenAI provider if the server's configuration arguments were mutated after initialization. This was caused by a lambda function capturing a reference to the mutable 'args' object instead of capturing the configuration values at creation time.
 2025-08-30 14:43:33 +02:00
yangdx
43f32e8d97 Bump api version to 0209 2025-08-29 19:42:06 +08:00
yangdx
f3989548b9 Fix MongoDB vector query embedding format compatibility 
* Convert numpy arrays to lists
* Ensure MongoDB compatibility
 2025-08-29 18:51:53 +08:00
yangdx
03d0fa3014 perf: add optional query_embedding parameter to avoid redundant embedding calls 2025-08-29 18:15:45 +08:00
yangdx
a923d378dd Remove deprecated ID-based filtering from vector storage queries 
- Remove ids param from QueryParam
- Simplify BaseVectorStorage.query signature
- Update all vector storage implementations
- Streamline PostgreSQL query templates
- Remove ID filtering from operate.py calls
 2025-08-29 17:06:48 +08:00
yangdx
d7e0701b63 Improve logging setup and add error prefixes for LLM functions 
- Move logger init to top of file
- Add console handler by default
- Prefix LLM errors with "[LLM func]"
- Update timeout log messages
- Comment out pypinyin success log
 2025-08-29 14:19:13 +08:00
yangdx
925e631a9a refac: Add robust time out handling for LLM request 2025-08-29 13:50:35 +08:00
yangdx
99e28e815b fix: prevent document processing failures from UTF-8 surrogate characters 
- Change sanitize_text_for_encoding to fail-fast instead of returning error placeholders
- Add strict UTF-8 cleaning pipeline to entity/relationship extraction
- Skip problematic entities/relationships instead of corrupting data

Fixes document processing crashes when encountering surrogate characters (U+D800-U+DFFF)
 2025-08-27 23:52:39 +08:00
yangdx
6a2a592224 Fix linting 2025-08-27 12:51:50 +08:00
yangdx
8a0d06e557 Restore default entity types 2025-08-27 12:51:18 +08:00
yangdx
28e07c89f9 Fix linting 2025-08-27 12:35:51 +08:00
yangdx
2ccc39de9a Fix language fallback in summarize error 2025-08-27 12:34:27 +08:00
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
LinkinPony
45da0385eb
Merge branch 'HKUDS:main' into main 2025-08-27 09:22:39 +08:00
Thibo Rosemplatt
c3aabfc251 Merge branch 'main' into entityTypesServerSupport 2025-08-26 21:48:20 +02:00
yangdx
c259b8f22c Update webui assets and bump aip verion to 0208 2025-08-26 23:05:00 +08:00
yangdx
d3623cc9ae fix: resolve infinite loop risk in _handle_entity_relation_summary 
- Ensure oversized descriptions are force-merged with subsequent ones
- Add len(current_list) <= 2 termination condition to guarantee convergence
- Implement token-based truncation in _summarize_descriptions to prevent overflow
 2025-08-26 21:58:31 +08:00
yangdx
e0a755e42c Refactor prompt instructions to emphasize depth and completeness 2025-08-26 18:28:57 +08:00
yangdx
79e0226b2b Refactor: move force_llm_summary_on_merge to global_config access 
- Remove parameter from function signature
- Access from global_config instead
- Improve code consistency
 2025-08-26 18:02:39 +08:00
yangdx
01a2c79f29 Standardize prompt formatting and section headers across templates 
- Remove hash delimiters
- Consistent section headers
- Add "Output:" labels
- Clean up example formatting
 2025-08-26 14:42:52 +08:00
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
LinkinPony
ff4c747a2a fix mismatch of 'error' and 'error_msg' in MongoDB 2025-08-26 10:43:56 +08:00