Download Results

[ascibisz]

Use Case

After running a packing, we should allow users to download cellpack output files from the cellpack client UI

Acceptance Criteria

• When a packing is successful, a download button should appear (or maybe it always exists, but now it's enabled)
• When the user selects the download button, a zip file of all of the generated output files related to that packing should be downloaded
• In order to build this zip file, we need to determine where all of these output files are getting generated during remote packings. We currently are only publishing the simularium files to s3, how can we do this for all of the files?
• NOTE: in early phases of investigating this ticket, assess how much work is entailed. If it's going to take a lot of work, we should decide if this is actually worth spending our time on before submission, or if we should work on other tasks

[rugeli]

clarifying questions:

• check output files upload logic: how many buckets are there? how are we handling different file types?
• should users be able to download individual files as well, or just the complete zip?

Output file types:

• json (results, analysis data, distance matrices)
• png (packing images, histograms, figures)
• simularium results(.simularium)
• grid data files (.dat)
• analysis reports (markdown) --- probably don't need for this feature

brainstorming:

• extend s3 upload logic to include all outputs
• output files tracking - how many files created and their locations in one job
• need another endpoint for this? like GET /jobs/{job_id}/download (potential storage/memory issue?)
• frontend work UI + error handling

current upload through:
.github/workflows/analyze.yml
aws s3 cp ./results s3://cellpack-results/${{ github.ref_name }}/ --recursive --acl public-read

files are organized by branch name: s3://cellpack-results/{branch_name}/