Currently the logger will generate log names based on hour minute second. This is slightly problematic as the dev server will create new logs on every file change. The dev server auto reloads on file changes.
We should consider making the load_ndcs module separately runnable from the web server itself. This would avoid the current process of commenting out part of the run.py module.
The productType table in the DB does not check for existence prior to adding potentially duplicate values. This allows for the same value to be entered multiple times if load_data.py is run.
Currently to load data into the database a file must be present in data/. This file name is currently hardcoded. I would suggest we come up with a solution to remove this dependency. This issue is likely dependent on #15
Wanted to create this issue mostly as food for thought. How do you guys feel about using a python package management lib, like pipenv or poetry. Both libraries are helpful to pin dependency versions and make dependencies easier to manage (install and update).
If these are kinda overkill at this point we may want to create a requirements.txt to pin dependency versions. This isnt a huge deal at this point but we may want to consider this. With the current setup when the docker image is built it will install the latest versions of all the python libraries. This may lead to some users having different versions of these libraries as they develop code for this.