mozexames-ml

Exams data extractor w/ Machine Learning.

Getting started

This app is using vanilla Python 3.10, leveraging VS Code's Dev Containers (a.k.a Remote Containers) + Docker Compose for its environment. For Machine Learning, this project uses the best pre-trained Portuguese Text Recognition model from tesseract-ocr, enabled by the pytesseract Python wrapper.

System requirements

Docker version >= 23.0 (Recommended to use the latest version if possible)
VS Code
That's it!

First time

Clone the project, cp .env.example .env then open it on your VS Code.
Install the ms-vscode-remote.remote-containers VS Code extension (if not already).
Open the Command Pallete CMD / CTRL + SHIFT + P then search and run Dev Containers: Reopen in Container.

This will start building the Dev Container using docker-compose.yml.

Once it's built, use the terminal within VS Code to execute other commands using ./run.

Ensure that VS Code has selected the correct Python interpreter within your new environment: /usr/local/bin/python.

Use ./run app to execute the main src/app.py script at anytime.

Note

Use ./run test run tests.

_{You can also append -k [name] to run tests matching the provided name in isolation}

Hint: The ./run app command essential does ./run python src/app.py

Warning

It is recommended to develop within a Dev Container as you'll be able to use the Python interpreter installed within your container, and correct IDE dependency suggestions and links via the pre-installed pylance extension.

As a side-note, you can manually build your the docker container using ./run build if you don't fancy using Dev Containers, and use the same ./run commands as you would, with the only difference being that your IDE will struggle to find the installed dependencies.

Development

Running the GUI

macOS

Download XQuartz, then install and reboot your machine.
Open XQuartz.app > Settings (from the menu bar) > Security > And tick both "Authenticate connections" & "Allow connections from other network clients".
1. Run xhost + localhost
2. Keep XQuartz open (at least whenever you're developing this codebase)
Install socat via brew install socat
Run socat TCP-LISTEN:6000,reuseaddr,fork UNIX-CLIENT:/tmp/.X11-unix/X0 on a separate terminal, and let it listen.
Any GUI interface from this codebase should now work.

Linux

On Linux, X11 should just work requiring further configurations.

Other commands

./run test to run tests.
./run python: Execute python commands. You can append -w to watch the file for changes (hot module reload). It uses breuleux/jurigged under the hood.
./run pip: Execute pip commands (./run pip install ... and ./run pip uninstall ... are disabled. You should use the managed ./run pip:install ... and ./run pip:uninstall ... instead.
- To add a new dependency, either manually add them to requirements.in, then ./run pip:install or pass the dependencies as its arguments if you don't want to manually modify requirements.in. The same applies to ./run pip:uninstall.
- This project uses pip-tools to manage its dependencies.
./run bash to enter bash within the container when outside of the Dev Container environment.