L.O.B.E.

LOBE is a recording client made specifically for TTS data collections. It supports multiple collections, single and multi-speaker, and can prompt sentences based on phonetic coverage.

Setup

Install the project

pip install .

For additional dependencies for development, run

pip install -e .[dev]

ffmpeg is required for audio recording and processing

Postgresql

LOBE uses a Postgresql database. A simple setup is provided in the docker-compose.yaml file. To start the database, run

docker-compose up -d

Note that the LOBE application is commented out in the docker-compose file as the application should be run via flask during development.

Configuration

A Flask instance path is expected to be set in the environment variable FLASK_INSTANCE_PATH. This path should contain a config.py file. Take a look at instance_folder_dev/config.py for an example.

Database migrations

LOBE uses Flask-Migrate for database migrations. Since the application relies on the FLASK_INSTANCE_PATH environment variable, it is useful to set it and export it for other commands.

export FLASK_INSTANCE_PATH=(pwd)/instance_folder_dev/

To create a new migration, run

flask --app lobe db migrate -m "message"

To apply the migration, run

flask --app lobe db upgrade

Creating initial data

To create the initial roles and configuration, run

flask --app lobe user add_default_roles
flask --app lobe configuration add_default
flask --app lobe user add  # this is interactive

Running the application

LOBE is a Flask application. To run it, run

flask --app lobe run  # for debugging add --debug and --reload


# Known bugs
- Playing audio in sessions does not work in Safari
- Greining, does not work. So it is not possible to verify recordings.