Sasafrass / straattaal

Use an RNN to generate Dutch slang and interpret your newly created words!

Straattaal

Ever wanted to generate new Dutch slang? Look no further! This application makes use of Recurrent Neural Networks trained on a comprehensive database of Dutch slang to generate new slang. Some words may immediately bring up a visceral feeling of what they mean, and this application allows you to provide the community with your interpretation of what the newly created slang means.

Examples

We have several pre-trained models available to generate novel words from multiple classes. The following is a selection of some of our favorite generated words per category. Can you come up with a better meaning for the slang words?

Straattaal

• rampa: Een grote ramp / A big disaster.
• joeko: Groot / Big.
• opjo: Cool / cool.
• daggie: Een mes / A knife.
• zittie: Irritant / Annoying.

Plaatsnamen

• Boschem
• De Heel
• Echterberg
• Zeernenbroek
• Schoosnijk

Dutch words in general

• tertokraat
• gebelenster
• besteerding
• verstrijf
• wulveroeking

How to use?

Instructions on how to run this application locally. If you don't have Postgres set up (with the correct environment variables), it should run a SQLite database with more or less the same functionality. This should be sufficient for testing purposes and making tweaks to the code. Everything in these blocks are considered terminal commands.

• git clone https://github.com/Sasafrass/straattaal
• Navigate to the directory where you cloned using your terminal, e.g. cd straattaal
• Use an environment manager to install the packages: either venv and pip ** python3 -m venv venv or python -m venv venv depending on your Python installation to create a new virtual environment within the working directory. ** source venv/bin/activate to activate your new virtual environment. ** python -m pip install -r requirements.txt or with python3 depending on your Python installation.
• Or, alternatively, conda: ** conda env create -f environment.yml ** conda activate straattaal
• flask db upgrade migrate your database to the latest schema.
• flask run
• You should now be able to access a fully working local copy at http://localhost:5000

Run with Docker

This project can also be run with Docker in the following way:

• Create a .env file in the parent directory with the following variables: POSTGRES_USER=<postgres_user> POSTGRES_DB=<postgres_db> POSTGRES_PASSWORD=<postgres_password> PG_PORT=<postgres_port> (standard port is 5432).
  • You should be able to set any combination of POSTGRES_USER, POSTGRES_DB, and POSTGRES_PASSWORD as it will use the same combination for both the environment variables in the server docker container hosting the Flask app as for the Postgres container.
• Then build and run docker-compose with docker-compose -f docker-compose-dev.yml up --build. This will build a full Docker Compose setup with the (latest) Dockerfile defined in your branch or folder.
• Database should be mounted to a volume, and thus data should be persisted between container restarts.
• Navigate to http://127.0.0.1:5000. Your local port 5000 should be forwarded to the Docker Compose network.