Type of Amazon forests satellite image. Service

FastAPI service for Amazon forests satellite image NN multilabel classification. I made accent on "industrial quality" code with next technologies:

• FastAPI
• Gitlab CI/CD (test, deploy, destroy)
• DVC
• Docker
• Unit & Integration tests with coverage report
• Linters (flake8 + wemake)

Disclaimers:

• the project was originally crated and maintained in GitLab local instance, some repo functionality may be unavailable
• the project was created by me and me only as part of the CVRocket professional development course
• this project is my first "industry grade" service for NN, for more advanced code and features please see car-plate projects

Location for manual test:

• https://amazon_forest_api.lydata.duckdns.org
• docs https://amazon_forest_api.lydata.duckdns.org/docs#/default/process_content_process_content_post

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 service

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

Build docker

• make build - you can define arguments DOCKER_TAG, DOCKER_IMAGE

Static analyse

• make lint - run linters

Tests

• make run_unit_tests - run unit tests
• make run_integration_tests - run integration tests
• make run_all_tests - run all tests