The way work requests Etag is computed in the UI / api should be the same. We use that to run the default embeddings, and marker heatmap after the data processing. That way when user goes to the data exploration they are already run and just need to be downloaded. Unfortunately that is broken as the test were updated separately and no longer match.
According to Martin, fixing it for the heatmap (or embeddings? @cosa65 ) should be easy enough to do without changing so that's a good start.
However, this will be brittle and easy to break. Some ideas were floated around like moving the Etag calculation to the api (which has downsides as it requires many more UI-api requests).
Background
The way work requests Etag is computed in the UI / api should be the same. We use that to run the default embeddings, and marker heatmap after the data processing. That way when user goes to the data exploration they are already run and just need to be downloaded. Unfortunately that is broken as the test were updated separately and no longer match.
More details and link to involved files here:
https://this-is-biomage.slack.com/archives/C05END63C03/p1689586867051569
Goal
The goal is to decide the best solution for this.
According to Martin, fixing it for the heatmap (or embeddings? @cosa65 ) should be easy enough to do without changing so that's a good start.
However, this will be brittle and easy to break. Some ideas were floated around like moving the Etag calculation to the api (which has downsides as it requires many more UI-api requests).