Add MongoDB as KV storage

examples/lightrag_ollama_neo4j_milvus_mongo_demo.py

@@ -10,6 +10,10 @@ if not os.path.exists(WORKING_DIR):
     os.mkdir(WORKING_DIR)
 print(f"WorkingDir: {WORKING_DIR}")
 
+# mongo
+os.environ["MONGO_URI"] = "mongodb://root:root@localhost:27017/"
+os.environ["MONGO_DATABASE"] = "LightRAG"
+
 # neo4j
 BATCH_SIZE_NODES = 500
 BATCH_SIZE_EDGES = 100
@@ -38,6 +42,7 @@ rag = LightRAG(
             texts=texts, embed_model="bge-m3:latest", host="http://127.0.0.1:11434"
         ),
     ),
+    kv_storage="MongoKVStorage",
     graph_storage="Neo4JStorage",
     vector_storage="MilvusVectorDBStorge",
 )

lightrag/kg/mongo_impl.py

@@ -0,0 +1,51 @@
+import os
+from tqdm.asyncio import tqdm as tqdm_async
+from dataclasses import dataclass
+from pymongo import MongoClient
+
+from lightrag.utils import logger
+
+from lightrag.base import BaseKVStorage
+
+
+@dataclass
+class MongoKVStorage(BaseKVStorage):
+    def __post_init__(self):
+        client = MongoClient(
+            os.environ.get("MONGO_URI", "mongodb://root:root@localhost:27017/")
+        )
+        database = client.get_database(os.environ.get("MONGO_DATABASE", "LightRAG"))
+        self._data = database.get_collection(self.namespace)
+        logger.info(f"Use MongoDB as KV {self.namespace}")
+
+    async def all_keys(self) -> list[str]:
+        return [x["_id"] for x in self._data.find({}, {"_id": 1})]
+
+    async def get_by_id(self, id):
+        return self._data.find_one({"_id": id})
+
+    async def get_by_ids(self, ids, fields=None):
+        if fields is None:
+            return list(self._data.find({"_id": {"$in": ids}}))
+        return list(
+            self._data.find(
+                {"_id": {"$in": ids}},
+                {field: 1 for field in fields},
+            )
+        )
+
+    async def filter_keys(self, data: list[str]) -> set[str]:
+        existing_ids = [
+            str(x["_id"]) for x in self._data.find({"_id": {"$in": data}}, {"_id": 1})
+        ]
+        return set([s for s in data if s not in existing_ids])
+
+    async def upsert(self, data: dict[str, dict]):
+        for k, v in tqdm_async(data.items(), desc="Upserting"):
+            self._data.update_one({"_id": k}, {"$set": v}, upsert=True)
+            data[k]["_id"] = k
+        return data
+
+    async def drop(self):
+        """ """
+        pass

lightrag/lightrag.py

@@ -46,6 +46,8 @@ from .kg.oracle_impl import OracleKVStorage, OracleGraphStorage, OracleVectorDBS
 
 from .kg.milvus_impl import MilvusVectorDBStorge
 
+from .kg.mongo_impl import MongoKVStorage
+
 # future KG integrations
 
 # from .kg.ArangoDB_impl import (
@@ -227,6 +229,7 @@ class LightRAG:
             # kv storage
             "JsonKVStorage": JsonKVStorage,
             "OracleKVStorage": OracleKVStorage,
+            "MongoKVStorage": MongoKVStorage,
             # vector storage
             "NanoVectorDBStorage": NanoVectorDBStorage,
             "OracleVectorDBStorage": OracleVectorDBStorage,

requirements.txt

@@ -12,6 +12,7 @@ ollama
 openai
 oracledb
 pymilvus
+pymongo
 pyvis
 tenacity
 # lmdeploy[all]