Merge pull request #143 from thinkthinking/main

[feat] Add API server implementation and endpoints

README.md

@@ -397,6 +397,125 @@ if __name__ == "__main__":
 
 </details>
 
+## API Server Implementation
+
+LightRAG also provides a FastAPI-based server implementation for RESTful API access to RAG operations. This allows you to run LightRAG as a service and interact with it through HTTP requests.
+
+### Setting up the API Server
+<details>
+<summary>Click to expand setup instructions</summary>
+
+1. First, ensure you have the required dependencies:
+```bash
+pip install fastapi uvicorn pydantic
+```
+
+2. Set up your environment variables:
+```bash
+export RAG_DIR="your_index_directory"  # Optional: Defaults to "index_default"
+```
+
+3. Run the API server:
+```bash
+python examples/lightrag_api_openai_compatible_demo.py
+```
+
+The server will start on `http://0.0.0.0:8020`.
+</details>
+
+### API Endpoints
+
+The API server provides the following endpoints:
+
+#### 1. Query Endpoint
+<details>
+<summary>Click to view Query endpoint details</summary>
+
+- **URL:** `/query`
+- **Method:** POST
+- **Body:**
+```json
+{
+    "query": "Your question here",
+    "mode": "hybrid"  // Can be "naive", "local", "global", or "hybrid"
+}
+```
+- **Example:**
+```bash
+curl -X POST "http://127.0.0.1:8020/query" \
+     -H "Content-Type: application/json" \
+     -d '{"query": "What are the main themes?", "mode": "hybrid"}'
+```
+</details>
+
+#### 2. Insert Text Endpoint
+<details>
+<summary>Click to view Insert Text endpoint details</summary>
+
+- **URL:** `/insert`
+- **Method:** POST
+- **Body:**
+```json
+{
+    "text": "Your text content here"
+}
+```
+- **Example:**
+```bash
+curl -X POST "http://127.0.0.1:8020/insert" \
+     -H "Content-Type: application/json" \
+     -d '{"text": "Content to be inserted into RAG"}'
+```
+</details>
+
+#### 3. Insert File Endpoint
+<details>
+<summary>Click to view Insert File endpoint details</summary>
+
+- **URL:** `/insert_file`
+- **Method:** POST
+- **Body:**
+```json
+{
+    "file_path": "path/to/your/file.txt"
+}
+```
+- **Example:**
+```bash
+curl -X POST "http://127.0.0.1:8020/insert_file" \
+     -H "Content-Type: application/json" \
+     -d '{"file_path": "./book.txt"}'
+```
+</details>
+
+#### 4. Health Check Endpoint
+<details>
+<summary>Click to view Health Check endpoint details</summary>
+
+- **URL:** `/health`
+- **Method:** GET
+- **Example:**
+```bash
+curl -X GET "http://127.0.0.1:8020/health"
+```
+</details>
+
+### Configuration
+
+The API server can be configured using environment variables:
+- `RAG_DIR`: Directory for storing the RAG index (default: "index_default")
+- API keys and base URLs should be configured in the code for your specific LLM and embedding model providers
+
+### Error Handling
+<details>
+<summary>Click to view error handling details</summary>
+
+The API includes comprehensive error handling:
+- File not found errors (404)
+- Processing errors (500)
+- Supports multiple file encodings (UTF-8 and GBK)
+</details>
+
 ## Evaluation
 ### Dataset
 The dataset used in LightRAG can be downloaded from [TommyChien/UltraDomain](https://huggingface.co/datasets/TommyChien/UltraDomain).

examples/lightrag_api_openai_compatible_demo.py

@@ -0,0 +1,164 @@
+from fastapi import FastAPI, HTTPException
+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}")
+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(
+        "gpt-4o-mini",
+        prompt,
+        system_prompt=system_prompt,
+        history_messages=history_messages,
+        api_key="YOUR_API_KEY",
+        base_url="YourURL/v1",
+        **kwargs,
+    )
+
+
+# Embedding function
+
+
+async def embedding_func(texts: list[str]) -> np.ndarray:
+    return await openai_embedding(
+        texts,
+        model="text-embedding-3-large",
+        api_key="YOUR_API_KEY",
+        base_url="YourURL/v1",
+    )
+
+
+# Initialize RAG instance
+rag = LightRAG(
+    working_dir=WORKING_DIR,
+    llm_model_func=llm_model_func,
+    embedding_func=EmbeddingFunc(
+        embedding_dim=3072, max_token_size=8192, func=embedding_func
+    ),
+)
+
+# Data models
+
+
+class QueryRequest(BaseModel):
+    query: str
+    mode: str = "hybrid"
+
+
+class InsertRequest(BaseModel):
+    text: str
+
+
+class InsertFileRequest(BaseModel):
+    file_path: 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))
+        )
+        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(request: InsertFileRequest):
+    try:
+        # Check if file exists
+        if not os.path.exists(request.file_path):
+            raise HTTPException(
+                status_code=404, detail=f"File not found: {request.file_path}"
+            )
+
+        # Read file content
+        try:
+            with open(request.file_path, "r", encoding="utf-8") as f:
+                content = f.read()
+        except UnicodeDecodeError:
+            # If UTF-8 decoding fails, try other encodings
+            with open(request.file_path, "r", encoding="gbk") as f:
+                content = f.read()
+
+        # 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 {request.file_path} 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"

lightrag/lightrag.py

@@ -85,9 +85,7 @@ class LightRAG:
 
     # LLM
     llm_model_func: callable = gpt_4o_mini_complete  # hf_model_complete#
-    llm_model_name: str = (
+    llm_model_name: str = "meta-llama/Llama-3.2-1B-Instruct"  #'meta-llama/Llama-3.2-1B'#'google/gemma-2-2b-it'
-        "meta-llama/Llama-3.2-1B-Instruct"  #'meta-llama/Llama-3.2-1B'#'google/gemma-2-2b-it'
-    )
     llm_model_max_token_size: int = 32768
     llm_model_max_async: int = 16
 

setup.py

@@ -1,6 +1,7 @@
 import setuptools
 from pathlib import Path
 
+
 # Reading the long description from README.md
 def read_long_description():
     try:
@@ -8,6 +9,7 @@ def read_long_description():
     except FileNotFoundError:
         return "A description of LightRAG is currently unavailable."
 
+
 # Retrieving metadata from __init__.py
 def retrieve_metadata():
     vars2find = ["__author__", "__version__", "__url__"]
@@ -17,18 +19,26 @@ def retrieve_metadata():
             for line in f.readlines():
                 for v in vars2find:
                     if line.startswith(v):
-                        line = line.replace(" ", "").replace('"', "").replace("'", "").strip()
+                        line = (
+                            line.replace(" ", "")
+                            .replace('"', "")
+                            .replace("'", "")
+                            .strip()
+                        )
                         vars2readme[v] = line.split("=")[1]
     except FileNotFoundError:
         raise FileNotFoundError("Metadata file './lightrag/__init__.py' not found.")
-    
+
     # Checking if all required variables are found
     missing_vars = [v for v in vars2find if v not in vars2readme]
     if missing_vars:
-        raise ValueError(f"Missing required metadata variables in __init__.py: {missing_vars}")
+        raise ValueError(
-    
+            f"Missing required metadata variables in __init__.py: {missing_vars}"
+        )
+
     return vars2readme
 
+
 # Reading dependencies from requirements.txt
 def read_requirements():
     deps = []
@@ -36,9 +46,12 @@ def read_requirements():
         with open("./requirements.txt") as f:
             deps = [line.strip() for line in f if line.strip()]
     except FileNotFoundError:
-        print("Warning: 'requirements.txt' not found. No dependencies will be installed.")
+        print(
+            "Warning: 'requirements.txt' not found. No dependencies will be installed."
+        )
     return deps
 
+
 metadata = retrieve_metadata()
 long_description = read_long_description()
 requirements = read_requirements()
@@ -51,7 +64,9 @@ setuptools.setup(
     description="LightRAG: Simple and Fast Retrieval-Augmented Generation",
     long_description=long_description,
     long_description_content_type="text/markdown",
-    packages=setuptools.find_packages(exclude=("tests*", "docs*")),  # Automatically find packages
+    packages=setuptools.find_packages(
+        exclude=("tests*", "docs*")
+    ),  # Automatically find packages
     classifiers=[
         "Development Status :: 4 - Beta",
         "Programming Language :: Python :: 3",
@@ -66,6 +81,8 @@ setuptools.setup(
     project_urls={  # Additional project metadata
         "Documentation": metadata.get("__url__", ""),
         "Source": metadata.get("__url__", ""),
-        "Tracker": f"{metadata.get('__url__', '')}/issues" if metadata.get("__url__") else ""
+        "Tracker": f"{metadata.get('__url__', '')}/issues"
+        if metadata.get("__url__")
+        else "",
     },
 )