Merge branch 'entityTypesServerSupport'

env.example

@@ -119,9 +119,14 @@ RERANK_BINDING=null
 ########################################
 ### Document processing configuration
 ########################################
-### Language: English, Chinese, French, German ...
-SUMMARY_LANGUAGE=English
 ENABLE_LLM_CACHE_FOR_EXTRACT=true
+
+### Document processing outpu language: English, Chinese, French, German ...
+SUMMARY_LANGUAGE=English
+
+### Entity types that the LLM will attempt to recognize
+# ENTITY_TYPES=["person", "organization", "location", "event", "concept"]
+
 ### Chunk size for document splitting, 500~1500 is recommended
 # CHUNK_SIZE=1200
 # CHUNK_OVERLAP_SIZE=100

lightrag/api/config.py

@@ -38,6 +38,7 @@ from lightrag.constants import (
     DEFAULT_OLLAMA_MODEL_NAME,
     DEFAULT_OLLAMA_MODEL_TAG,
     DEFAULT_RERANK_BINDING,
+    DEFAULT_ENTITY_TYPES,
 )
 
 # use the .env that is inside the current folder
@@ -351,6 +352,7 @@ def parse_args() -> argparse.Namespace:
     # Add environment variables that were previously read directly
     args.cors_origins = get_env_value("CORS_ORIGINS", "*")
     args.summary_language = get_env_value("SUMMARY_LANGUAGE", DEFAULT_SUMMARY_LANGUAGE)
+    args.entity_types = get_env_value("ENTITY_TYPES", DEFAULT_ENTITY_TYPES)
     args.whitelist_paths = get_env_value("WHITELIST_PATHS", "/health,/api/*")
 
     # For JWT Auth

lightrag/api/lightrag_server.py

@@ -499,7 +499,10 @@ def create_app(args):
             rerank_model_func=rerank_model_func,
             max_parallel_insert=args.max_parallel_insert,
             max_graph_nodes=args.max_graph_nodes,
-            addon_params={"language": args.summary_language},
+            addon_params={
+                "language": args.summary_language,
+                "entity_types": args.entity_types,
+            },
             ollama_server_infos=ollama_server_infos,
         )
     else:  # azure_openai
@@ -526,7 +529,10 @@ def create_app(args):
             rerank_model_func=rerank_model_func,
             max_parallel_insert=args.max_parallel_insert,
             max_graph_nodes=args.max_graph_nodes,
-            addon_params={"language": args.summary_language},
+            addon_params={
+                "language": args.summary_language,
+                "entity_types": args.entity_types,
+            },
             ollama_server_infos=ollama_server_infos,
         )
 

lightrag/api/utils_api.py

@@ -264,6 +264,8 @@ def display_splash_screen(args: argparse.Namespace) -> None:
     ASCIIColors.magenta("\n⚙️ RAG Configuration:")
     ASCIIColors.white("    ├─ Summary Language: ", end="")
     ASCIIColors.yellow(f"{args.summary_language}")
+    ASCIIColors.white("    ├─ Entity Types: ", end="")
+    ASCIIColors.yellow(f"{args.entity_types}")
     ASCIIColors.white("    ├─ Max Parallel Insert: ", end="")
     ASCIIColors.yellow(f"{args.max_parallel_insert}")
     ASCIIColors.white("    ├─ Chunk Size: ", end="")

lightrag/constants.py

@@ -11,7 +11,7 @@ DEFAULT_WOKERS = 2
 DEFAULT_MAX_GRAPH_NODES = 1000
 
 # Default values for extraction settings
-DEFAULT_SUMMARY_LANGUAGE = "English"  # Default language for summaries
+DEFAULT_SUMMARY_LANGUAGE = "English"  # Default language for document processing
 DEFAULT_MAX_GLEANING = 1
 
 # Number of description fragments to trigger LLM summary
@@ -22,6 +22,16 @@ DEFAULT_SUMMARY_MAX_TOKENS = 1200
 DEFAULT_SUMMARY_LENGTH_RECOMMENDED = 600
 # Maximum token size sent to LLM for summary
 DEFAULT_SUMMARY_CONTEXT_SIZE = 12000
+# Default entities to extract if ENTITY_TYPES is not specified in .env
+DEFAULT_ENTITY_TYPES = [
+    "organization",
+    "person",
+    "geo",
+    "event",
+    "category",
+    "Equipment",
+    "Location",
+]
 
 # Separator for graph fields
 GRAPH_FIELD_SEP = "<SEP>"

lightrag/lightrag.py

@@ -39,6 +39,8 @@ from lightrag.constants import (
     DEFAULT_MAX_ASYNC,
     DEFAULT_MAX_PARALLEL_INSERT,
     DEFAULT_MAX_GRAPH_NODES,
+    DEFAULT_ENTITY_TYPES,
+    DEFAULT_SUMMARY_LANGUAGE,
 )
 from lightrag.utils import get_env_value
 
@@ -347,7 +349,10 @@ class LightRAG:
 
     addon_params: dict[str, Any] = field(
         default_factory=lambda: {
-            "language": get_env_value("SUMMARY_LANGUAGE", "English", str)
+            "language": get_env_value(
+                "SUMMARY_LANGUAGE", DEFAULT_SUMMARY_LANGUAGE, str
+            ),
+            "entity_types": get_env_value("ENTITY_TYPES", DEFAULT_ENTITY_TYPES, list),
         }
     )
 

lightrag/operate.py

@@ -47,6 +47,8 @@ from .constants import (
     DEFAULT_MAX_TOTAL_TOKENS,
     DEFAULT_RELATED_CHUNK_NUMBER,
     DEFAULT_KG_CHUNK_PICK_METHOD,
+    DEFAULT_ENTITY_TYPES,
+    DEFAULT_SUMMARY_LANGUAGE,
 )
 from .kg.shared_storage import get_storage_keyed_lock
 import time
@@ -1650,11 +1652,9 @@ async def extract_entities(
 
     ordered_chunks = list(chunks.items())
     # add language and example number params to prompt
-    language = global_config["addon_params"].get(
-        "language", PROMPTS["DEFAULT_LANGUAGE"]
-    )
+    language = global_config["addon_params"].get("language", DEFAULT_SUMMARY_LANGUAGE)
     entity_types = global_config["addon_params"].get(
-        "entity_types", PROMPTS["DEFAULT_ENTITY_TYPES"]
+        "entity_types", DEFAULT_ENTITY_TYPES
     )
     example_number = global_config["addon_params"].get("example_number", None)
     if example_number and example_number < len(PROMPTS["entity_extraction_examples"]):

lightrag/prompt.py

@@ -4,13 +4,10 @@ from typing import Any
 
 PROMPTS: dict[str, Any] = {}
 
-PROMPTS["DEFAULT_LANGUAGE"] = "English"
 PROMPTS["DEFAULT_TUPLE_DELIMITER"] = "<|>"
 PROMPTS["DEFAULT_RECORD_DELIMITER"] = "##"
 PROMPTS["DEFAULT_COMPLETION_DELIMITER"] = "<|COMPLETE|>"
 
-PROMPTS["DEFAULT_ENTITY_TYPES"] = ["organization", "person", "geo", "event", "category"]
-
 PROMPTS["DEFAULT_USER_PROMPT"] = "n/a"
 
 PROMPTS["entity_extraction"] = """---Goal---