Feature/firebase lookup

[rugeli]

Problem

closes #441

Solution

• replaced uuid-based job_id with dedup_hash(SHA256) for both params: recipe(the string path) and json_recipe(dict recipe data)
• added firebase lookup to check for existing results before packing, return cached result_path if found
• calculated dedup_hash via RecipeLoader.get_dedup_hash() after recipe normalization to make sure consistent hashing regardless of source (local_path, firebase_path, json body)

note: the previous approach hashes the normalized recipe, so the same semantic recipes from different sources(local_path, firebase_path, json body) would get the same hash. After chatting with @ascibisz, we simplified this to hash only the json body, aligning with server's planned use case

Type of change

• New feature (non-breaking change which adds functionality)

Steps to Verify:

1. run a packing job via the server and verify the document is created in job_status in firebase
2. run the exact same job again, no new document should be created, and the result_path in the web response should reference the cached result

[github-actions]

Packing analysis report

Analysis for packing results located at cellpack/tests/outputs/test_spheres/spheresSST

Ingredient name	Encapsulating radius	Average number packed
ext_A	25	236.0

Packing image

Distance analysis

Expected minimum distance: 50.00
Actual minimum distance: 50.01

Ingredient key	Pairwise distance distribution
ext_A

[ascibisz]

When I tried running this locally, I got the error ERROR | DBRecipeHandler:647 | upload_packing_results_workflow() | 'AWSHandler' object has no attribute 'create_timestamp', which I think traces back to the fact that we're now calling upload_job_status here instead of update_outputs_directory here, so we need to add a check into upload_job_status similiar to the one we had in update_outputs_directory since self.db is currently a AWSHandler instance.

So I think we want to add something like

if not self.db or self.db.s3_client:
    # switch to firebase handler to update job status
    handler = DATABASE_IDS.handlers().get("firebase")
    initialized_db = handler(default_db="staging")

to upload_job_status

[ascibisz]

I think we want to move result_path out to an if statement, like we do for outputs_directory, so we don't overwrite it's value in firebase to None

[rugeli]

good catch!

[ascibisz]

I think we need to add the line setting this job ID back in since we try to read it in simularium_helper, or we need to update simularium_helper to have dedup_hash passed to post_and_open_file

[rugeli]

Thanks for catching that! I refactored things to pass dedup_hash through env, seems also helped address the TODO in simulairum_helper too. let me know what do you think about this approach, here is the commit