broadinstitute/genetic-prevalence-estimator

Keep track of list of variants to not include in calculations

Closed this issue · 1 comments

Some overlap with, and possible superceding of #124

Per discussion in meeting on Jan 10, 2024, support should be added to add variant IDs to a list that never gets included in calculations.

The use case is that when a user knows that a certain set of variants will never be included, but also does some experimenting with de-selecting another set, when they go back to re-enable all variants, both sets get added. There could be a known set of variant IDs that the variant list keeps track of, that will not get checked and included in calculations.

Just to add this information in writing:

I think a good way to start to go about this, particularly with the backend work, is to follow how to a piece of information from the frontend makes it to the backend and to the DB.

Using the network tab of the dev tools of a browser, you can see what endpoint gets hit when a box is checked or unchecked, from there you can use the endpoint to find what View is used, what View is used to find was Serializer is used, and what Serializer is used to see what Model Fields are interacted with. This should be a useful guide of a piece of data making its way from a request from the frontend, to being stored in the DB. Using that knowledge can help in adding the plumbing to interact with a new field from the frontend, and once we've confirmed the frontend correctly gets and changes information in the DB, we can make the frontend changes to the calculations.

In general, the Django docs are pretty useful, and Django is widely used enough that there is lots of good information about on various forums (and that's now probably been used in training sets on LLMs, though imo use LLMs advice with a grain of salt and a bit of caution).

Definitely reach out at any point with questions!