Car plates project. Additional example how to use API service in Telegram bot

This is the project for car plate OCR recognition, which include:

1. Neural network segmentation model for car plate area with number selection (part 1/3)
2. Neural network OCR model for plate character recognition (part 2/3)
3. API service for these two models (part 3/3)
4. Additional example how to use API service in Telegram bot

A decentralized parking app that automatically pays for parking using geolocation and plate recognition was created using the API service. The app was developed during the TON x ETH Belgrade hackathon. For more information, check out the project's presentation and tg bot.

Used technologies:

• Aiogram (for telegram bot)
• Asyncio
• Requests, for communication with API
• CI/CD (test, deploy, destroy)
• Linters (flake8 + wemake, pylint)

Disclaimers:

• the project was originally crated and maintained in GitLab local instance, some repo functionality may be unavailable
• the project was created by my team "LogicYield" for TON х ETH Belgrade Hackathon
• according to time limitation main code include in main.py file

Setup of environment

First, create and activate venv: bash python3 -m venv venv . venv/bin/activate

Next, install dependencies: bash make install

Commands

Preparation

• make install - install python dependencies

Run telegram bot

• make run_app - run servie. You can define argument APP_PORT

Static analyse

• make lint - run linters

Build docker

• make build - you can define arguments DOCKER_TAG, DOCKER_IMAGE

Run app in docker

• make docker_run - you can define arguments DOCKER_NAME, DOCKER_TAG, DOCKER_IMAGE