Docker:

• To build locally sudo docker build -t alice-flask .

• Get the image id from this sudo docker image ls

• To run on port 8080 sudo docker run -d -p 8080:8080 <IMAGE_ID>

To test (also in pytests in Docker)

• Open a python console with poetry in a new terminal: poetry run python In the console:

  • Post request requests.post('http://10.0.0.78:8080/post', '{"x_value": 1.88, "calculation_type": "red", "y_translation": 7}').text

  • Get request to show all table requests.get('http://10.0.0.78:8080/calculations').text

  • Get request with id requests.get('http://10.0.0.78:8080/calculations/1').text

  • Get request with since query requests.get('http://10.0.0.78:8080/calculations?since=100').text

Notes

• The poetry link in the test is deprecated. The poetry documentation even says do not install poetry this way https://python-poetry.org/docs/.
• poetry init, go through options. Creates pyproject.toml poetry is basically like conda. A virtual environment to manage packages.
• 'poetry shell' opens the environment (like conda activate my_env)
• poetry add --> Add package to environment, then can do poetry run python app.py and it will work.
• Container is a running image, e.g alice-app.
• Image is created that gets everything ready to go. When we do docker run .. it becomes a container and the app is live.
• Making a SQL table in memory in one script means it could not be accessed by app.py script.
• Flask accesses the requests to the database using threads.
• We need to put the db on the disc so that is can be accessed by multiple threads.
• Had to add id argument to start_calculation so I can update the table.
• json was fiddly, going between strings and dictionaries and json
• Choice between datetime.time() and time.time() for on_complete and on_error was tricky. datetime is pretty but need time in second for since query.
• In dockerfile, could remake poetry and add everything again or I have just copied my existing .toml file.