ispovesti.ml-backend

This is the source code of the backend service for ispovesti.ml

Installation

First, acquire the source code, either by cloning this repository or downloading it directly

Cloning:

git clone [email protected]:dojitza/ispovesti.ml-backend.git

Two installation methods are provided: You can use docker compose or manually install the components.

A. Docker-compose

Use the provided docker-compose file to build and start the backend automatically. Make sure to have docker-compose installed prior to running this command.

$ cd ispovesti.ml-backend
$ docker-compose up

B. Manual

1. Setup the virtual environment, and install the requirements.

$ cd ispovesti.ml-backend
$ python3 -m venv venv
$ source venv/bin/activate
$ pip install -r requirements.txt

2. Install the rabbitmq broker and make sure its service is started (you may wish to enable it as well) . This assumes you use apt as your package manager and systemd as your init daemon

$ sudo apt install rabbitmq-server
$ sudo systemctl start rabbitmq
$ sudo systemctl enable rabbitmq

3. Download and unpack the model from the corresponding version github assets (the following link is for v2.0) https://github.com/dojitza/ispovesti.ml-backend/releases/download/v2.0/trained_model.tar
   After unpacking, your tree should look like this:

├── LICENSE.md
├── README.md
├── checkpoint
│   └── run1
│       ├── checkpoint
│       ├── counter
│       ├── encoder.json
│       ├── events.out.tfevents.1599479304.22233ed79ca3
│       ├── hparams.json
│       ├── model-39563.data-00000-of-00001
│       ├── model-39563.index
│       ├── model-39563.meta
│       └── vocab.bpe
├── constants.py
├── dailyReset.py
├── db.py
...

Launching the service:

You can use gunicorn, or any other wsgi server.

cd ispovesti.ml-backend
pip3 install gunicorn
gunicorn main:app --bind 0.0.0.0:8080

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

GNU GPLv3