增强文本生成

Building from Source

Before building the binary, clone the RAG server repository and navigate into the root of the repo:

git clone https://github.com/pgedge/pgedge-rag-server.git
cd pgedge-rag-server

Build the pgEdge RAG server binary with the command; the binary is created in the bin directory:

make build

After installation, verify the tool is working:

pgedge-rag-server version

You can also access online help after building RAG server:

pgedge-rag-server help

Quick Start - Using the RAG Server

To use the pgEdge RAG Server, you must:

1. Build the pgedge-rag-server binary.
2. Create a configuration file that specifies details used by the RAG server.
3. Invoke pgedge-rag-server.

Before installing pgEdge RAG Server, you should install or obtain:

• Go 1.22 or later
• PostgreSQL 14 or later, with pgvector installed
• API keys for your chosen LLM providers

API Usage Reference

The online documentation contains detailed information about using the API, and allows you to try the API in a browser.

To List Available Pipelines

curl http://localhost:8080/v1/pipelines

To Query a Pipeline

curl -X POST http://localhost:8080/v1/pipelines/my-docs \
  -H "Content-Type: application/json" \
  -d '{"query": "How do I configure replication?"}'

To Query with Streaming

curl -X POST http://localhost:8080/v1/pipelines/my-docs \
  -H "Content-Type: application/json" \
  -d '{"query": "How do I configure replication?", "stream": true}'

Creating a Configuration File

Create a configuration file that specifies server connection details and other properties; (see the online documentation for complete details. The default name of the file is pgedge-rag-server.yaml; when invoked, the server searches for configuration file in:

1. /etc/pgedge/pgedge-rag-server.yaml
2. the directory that contains the pgedge-rag-server binary.

You can optionally use the -config option on the command line to specify the complete path to a custom location for the configuration file.

The following sample demonstrates a minimal configuration:

server:
  listen_address: "0.0.0.0"
  port: 8080

pipelines:
  - name: "my-docs"
    description: "Search my documentation"
    database:
      host: "localhost"
      port: 5432
      database: "mydb"
    tables:
      - table: "documents"
        text_column: "content"
        vector_column: "embedding"
    embedding_llm:
      provider: "openai"
      model: "text-embedding-3-small"
    rag_llm:
      provider: "anthropic"
      model: "claude-sonnet-4-20250514"