The objective of this project is to develop a comprehensive solution for identifying suitable songs based on user-provided inputs, including videos, textual descriptions, or music.

• Python 3.10
• Dependencies listed in requirements.txt

1. Clone the repository:

   git clone https://github.com/alldbi/MultiMediaAI

2. Install dependencies:

   Note: Python 3.10 is required for the installation of the dependencies. If you are using a different version of Python, you can create a virtual environment with Python 3.10 and install the dependencies there.

   cd MultiMediaAI
   pip install -r requirements.txt
   sudo apt-get install poppler-utils

3. Connect a device for YouTube:

   If it's your first time running the application, you need to connect a device for YouTube. Run the following script:

   python setup_yt.py

   And follow the instructions. You should see a message Setup finished successfully! if everything went well.

   If you're unsure whether the device has already been set up, you can still run this script. It won't ask you to connect the device again if it's already set up.

4. Setup the Suno API:

   Follow the instructions in the Suno API Setup Documentation to set up the Suno API.

The Gradio app provides an interactive UI for the backend functionalities.

python -m src.app

Access the application interface via your web browser at http://127.0.0.1:5002.

Port 5002 is used by default and can be changed in the configs/config.py file.

You can find additional information about settings and the UI in the Gradio App Documentation.

The FastAPI app provides API endpoints for backend functionalities.

python -m src.api

Application will run on port 5011.

Port 5011 is used by default and can be changed in the configs/config.py file.

You can find additional information about settings and the API endpoints in the API Documentation.

To ensure compatibility with the latest changes in YouTube's infrastructure, it is important to regularly update the PyTubeFix library. PyTubeFix is a Python library that fixes issues with downloading audio from YouTube.

To update the PyTubeFix library, you can use the following command:

pip install --upgrade pytubefix

By keeping the PyTubeFix library up to date, you can ensure that your application continues to work seamlessly with YouTube's latest changes and improvements.

The project includes several configuration files that control various aspects of the application. These files are located in the configs directory.

More information about the configuration files can be found in the Configuration Files Documentation.

You need to provide your API keys in a .env file in the root directory of the project. Also, provide credentials for the AWS S3 bucket where the generated music files will be stored.

Contents of the .env file should look like this:

OPENAI_API_KEY = "sk-XXXXXXXX"
CYANITE_ACCESS_TOKEN = "Bearer XXX"
GRADIO_PASSWORD = "password"

AWS_ACCESS_KEY_ID = ...
AWS_SECRET_ACCESS_KEY = ...
AWS_REGION = ...
AWS_BUCKET = ...

ASSISTANT_ID = ...
ASSISTANT_KNOWLEDGE_FILE_ID = ... # you can find file id in the project settings, storage section

Note: Generated music will be stored in folder SUNO_S3_FOLDER from configs/config.py in the specified S3 bucket.# Sff-backend