Refactor uploading system deduplication using hash

[rugeli]

Problem

What is the problem this work solves, including
closes #316

Solution

What I/we did to solve this problem

• used hashlib to generate a unique hashes for each doc (any changes will produce a new hash)

• removed the should_write method from each doc class

• updated unit tests
  with @ascibisz

• introduced new field dedup_hash

Type of change

Please delete options that are not relevant.

• Refactor (non-breaking change which adds functionality)

Steps to Verify:

1. Upload a recipe to your firebase dev db
2. Re-upload the same recipe to confirm the hash prevents duplicates
3. edit recipe's name and a nested data field, then upload again to test if new doc and new recipe are detected
4. run pytest

[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

[rugeli]

TODO: troubleshoot upload_composition to detect any changes in inner compositions and create new top level compositions.

i.e. updating the count of A should create a new comp A and a new comp space. These new compositions should be linked to the corresponding recipe as references.

 "composition": {
        "space": {
            "regions": {
                "interior": [
                    "A"
                ]
            }
        },
        "A": {
            "object": "sphere_25",
            "count": 1
        }
    }

[rugeli]

TODO: troubleshoot upload_composition to detect any changes in inner compositions and create new top level compositions.

i.e. updating the count of A should create a new comp A and a new comp space. These new compositions should be linked to the corresponding recipe as references.

 "composition": {
        "space": {
            "regions": {
                "interior": [
                    "A"
                ]
            }
        },
        "A": {
            "object": "sphere_25",
            "count": 1
        }
    }

Solution:
We've implemented topological sorting to make sure that the compositions are processed in the correct dependency order, from innermost to outermost, with proper firebase reference replacement. So in the case above, a field in A is modified, the system will detect the change and create a new A, a new space and a new recipe.