File size: 4,890 Bytes
f7ab812
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
from fastapi import FastAPI, HTTPException, File, UploadFile
from pydantic import BaseModel
import os
from lightrag import LightRAG, QueryParam
from lightrag.llm import openai_complete_if_cache, openai_embedding
from lightrag.utils import EmbeddingFunc
import numpy as np
from typing import Optional
import asyncio
import nest_asyncio

# Apply nest_asyncio to solve event loop issues
nest_asyncio.apply()

DEFAULT_RAG_DIR = "index_default"
app = FastAPI(title="LightRAG API", description="API for RAG operations")

# Configure working directory
WORKING_DIR = os.environ.get("RAG_DIR", f"{DEFAULT_RAG_DIR}")
print(f"WORKING_DIR: {WORKING_DIR}")
LLM_MODEL = os.environ.get("LLM_MODEL", "gpt-4o-mini")
print(f"LLM_MODEL: {LLM_MODEL}")
EMBEDDING_MODEL = os.environ.get("EMBEDDING_MODEL", "text-embedding-3-large")
print(f"EMBEDDING_MODEL: {EMBEDDING_MODEL}")
EMBEDDING_MAX_TOKEN_SIZE = int(os.environ.get("EMBEDDING_MAX_TOKEN_SIZE", 8192))
print(f"EMBEDDING_MAX_TOKEN_SIZE: {EMBEDDING_MAX_TOKEN_SIZE}")

if not os.path.exists(WORKING_DIR):
    os.mkdir(WORKING_DIR)


# LLM model function


async def llm_model_func(
    prompt, system_prompt=None, history_messages=[], **kwargs
) -> str:
    return await openai_complete_if_cache(
        LLM_MODEL,
        prompt,
        system_prompt=system_prompt,
        history_messages=history_messages,
        **kwargs,
    )


# Embedding function


async def embedding_func(texts: list[str]) -> np.ndarray:
    return await openai_embedding(
        texts,
        model=EMBEDDING_MODEL,
    )


async def get_embedding_dim():
    test_text = ["This is a test sentence."]
    embedding = await embedding_func(test_text)
    embedding_dim = embedding.shape[1]
    print(f"{embedding_dim=}")
    return embedding_dim


# Initialize RAG instance
rag = LightRAG(
    working_dir=WORKING_DIR,
    llm_model_func=llm_model_func,
    embedding_func=EmbeddingFunc(
        embedding_dim=asyncio.run(get_embedding_dim()),
        max_token_size=EMBEDDING_MAX_TOKEN_SIZE,
        func=embedding_func,
    ),
)


# Data models


class QueryRequest(BaseModel):
    query: str
    mode: str = "hybrid"
    only_need_context: bool = False


class InsertRequest(BaseModel):
    text: str


class Response(BaseModel):
    status: str
    data: Optional[str] = None
    message: Optional[str] = None


# API routes


@app.post("/query", response_model=Response)
async def query_endpoint(request: QueryRequest):
    try:
        loop = asyncio.get_event_loop()
        result = await loop.run_in_executor(
            None,
            lambda: rag.query(
                request.query,
                param=QueryParam(
                    mode=request.mode, only_need_context=request.only_need_context
                ),
            ),
        )
        return Response(status="success", data=result)
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))


@app.post("/insert", response_model=Response)
async def insert_endpoint(request: InsertRequest):
    try:
        loop = asyncio.get_event_loop()
        await loop.run_in_executor(None, lambda: rag.insert(request.text))
        return Response(status="success", message="Text inserted successfully")
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))


@app.post("/insert_file", response_model=Response)
async def insert_file(file: UploadFile = File(...)):
    try:
        file_content = await file.read()
        # Read file content
        try:
            content = file_content.decode("utf-8")
        except UnicodeDecodeError:
            # If UTF-8 decoding fails, try other encodings
            content = file_content.decode("gbk")
        # Insert file content
        loop = asyncio.get_event_loop()
        await loop.run_in_executor(None, lambda: rag.insert(content))

        return Response(
            status="success",
            message=f"File content from {file.filename} inserted successfully",
        )
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))


@app.get("/health")
async def health_check():
    return {"status": "healthy"}


if __name__ == "__main__":
    import uvicorn

    uvicorn.run(app, host="0.0.0.0", port=8020)

# Usage example
# To run the server, use the following command in your terminal:
# python lightrag_api_openai_compatible_demo.py

# Example requests:
# 1. Query:
# curl -X POST "http://127.0.0.1:8020/query" -H "Content-Type: application/json" -d '{"query": "your query here", "mode": "hybrid"}'

# 2. Insert text:
# curl -X POST "http://127.0.0.1:8020/insert" -H "Content-Type: application/json" -d '{"text": "your text here"}'

# 3. Insert file:
# curl -X POST "http://127.0.0.1:8020/insert_file" -H "Content-Type: application/json" -d '{"file_path": "path/to/your/file.txt"}'

# 4. Health check:
# curl -X GET "http://127.0.0.1:8020/health"