MCP

🚀 File Search and Speech Transcription with MCP

This project leverages the Multi-Channel Processing (MCP) framework to implement file search and speech transcription functionalities. It provides REST APIs through an HTTP server, allowing users to search for files and transcribe audio easily.

🚀 Quick Start

📦 Installation

First, install the project dependencies:

npm install
pip install -r requirements.txt  # Make sure you've created the requirements.txt file

🔨 Compilation and Build

Compile the TypeScript code to generate JavaScript files:

npm run build

🚀 Start the Services

File Search Server (Direct Implementation)

Run src/index.ts:

ts-node src/index.ts

HTTP File Search Server

Run main.py:

python main.py

By default, the server will start on port 8080.

Speech Transcription Server (Direct Implementation)

Run src/whisper-index.ts:

ts-node src/whisper-index.ts

HTTP Speech Transcription Server

Run whisper_server.py:

python whisper_server.py

By default, the server will start on port 8081.

✨ Features

• File Search: Provide a REST API through an HTTP server, allowing users to search for files in a specified path by sending requests.
• Speech Transcription: Use the fastspeech library to convert audio to text and provide detailed transcription information.

📦 Project Structure

The project directory structure is as follows:

project/
├── src/
│   ├── index.ts        # MCP server implementation for file search (direct implementation)
│   ├── index-http.ts    # MCP proxy implementation for file search on the HTTP server
│   ├── whisper-index.ts # MCP server implementation for speech transcription (direct implementation)
│   └── whisper-index-http.ts # MCP proxy implementation for speech transcription on the HTTP server
├── build/
│   ├── index.js        # Compiled JavaScript file for the MCP server for file search
│   ├── index-http.js    # Compiled JavaScript file for the MCP proxy for file search on the HTTP server
│   ├── whisper-index.js # Compiled JavaScript file for the MCP server for speech transcription
│   └── whisper-index-http.js # Compiled JavaScript file for the MCP proxy for speech transcription on the HTTP server
├── tsconfig.json        # TypeScript configuration file
├── package.json         # Project dependencies and description
├── main.py              # Python implementation of the HTTP server for file search
└── whisper_server.py    # Python implementation of the HTTP server for speech transcription

💻 Usage Examples

📂 File Search API

Send a GET request to http://localhost:8080/search?path=/target/directory to get all files in the specified path.

Request Example:

curl http://localhost:8080/search?path=/home/user/documents

Response Example:

{
  "files": [
    "/home/user/documents/report.pdf",
    "/home/user/documents/data.csv"
  ],
  "directories": [
    "/home/user/documents/subfolder"
  ]
}

🎤 Speech Transcription API

Send a POST request to http://localhost:8081/transcribe and include audio data in the request body.

Request Example:

curl -X POST http://localhost:8081/transcribe \
     --header "Content-Type: application/json" \
     --data '{"audio":"JVBERi0xLjUK...","language":"zh-CN"}'

Response Example:

{
  "text": "这是一个语音转写示例。",
  "segments": [
    {
      "start": 0.0,
      "end": 2.5,
      "text": "这是"
    },
    {
      "start": 2.5,
      "end": 4.0,
      "text": "一个语音转写示例。"
    }
  ],
  "language": "zh-CN",
  "language_probability": 0.98
}

🛠️ Troubleshooting

⚠️ Important Note

• If you encounter a "connection not found" error, make sure to restart VS Code after modifying the MCP settings.
• If the server doesn't respond, check if the MCP proxy path correctly points to the compiled JavaScript file.
• When using fastspeech with GPU acceleration, ensure that the necessary graphics card drivers and computing libraries (such as CUDA) are installed.

📚 Appendix

🐳 Deployment with Docker

You can package the project as a Docker image for production use. Create a Dockerfile as follows:

FROM node:16-alpine AS builder
WORKDIR /app
COPY package*.json ./
RUN npm install --production

FROM python:3.8-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install -r requirements.txt
COPY --from=builder /app/ .

# Set the default Python path
ENV PATH=/usr/local/bin:$PATH

EXPOSE 8080 8081
CMD ["python", "main.py", "--port", "8080"]

Then build and run the image:

docker build -t mcp .
docker run -p 8080:8080 -p 8081:8081 mcp

📁 Example Project Structure

Ensure that the project root directory contains the following files and subdirectories:

project/
├── src/
│   ├── index.ts
│   └── whisper-index.ts
├── build/
│   ├── index.js
│   └── whisper-index.js
├── package.json
├── tsconfig.json
├── main.py
└── whisper_server.py

📦 Dependency Management

Make sure you have installed the Node.js and Python development environments and the project dependencies.

Node.js Dependencies:

{
  "name": "mcp",
  "version": "1.0.0",
  "description": "Multi-Channel Processing for file search and speech transcription.",
  "scripts": {
    "build": "tsc -p tsconfig.json"
  },
  "dependencies": {
    "typescript": "^4.5.2",
    "fastspeech": "^0.13.0"
  }
}

Python Dependencies:

numpy
flask
pydub

🤝 Contribution Guidelines

Contributions are welcome! Please follow these steps:

1. Create a feature branch: git checkout -b feature/your-feature
2. Commit your changes: git add . && git commit -m 'Your commit message'
3. Push to the remote repository: git push origin feature/your-feature
4. Submit a pull request

📄 License

This project is licensed under the MIT License. For details, please refer to the LICENSE file.