Compartment names should be parameterizable

[shntnu]

It's possible this is already the case, but @pamzgt wasn't able to find it

[gwaybio]

the default is ["cells", "cytoplasm", and "nuclei"], but compartments are indeed parameterizable:

pycytominer/pycytominer/aggregate.py

Line 30 in ac3ef68

compartments=["cells", "cytoplasm", "nuclei"],

It is possible they are not parameterized somewhere, but if so, we should rectify.

[gwaybio]

ah, I just realized that pycytominer will throw an error for non-canonical compartments:

pycytominer/pycytominer/cyto_utils/util.py

Line 15 in ac3ef68

def check_compartments(compartments):

We should probably update this to throw a warning instead.

[pamzgt]

My bad. I was taking a very quick look and did not realize that compartments referred to the table names in the sqlite file.

[pamzgt]

To give a bit of context, I was trying to feed an sqlite db with the following tables: 'Per_Image', 'Per_Cells', 'Per_Cytoplasm', 'Per_Nuclei'.

passing compartments =["Per_Cells", "Per_Cytoplasm", "Per_Nuclei"] results in an error: compartment not supported, use one of ['cells', 'cytoplasm', 'nuclei']. Also, the image table should be called 'Image' and can not be parameterized.

At that moment, I did not know about cytominer-database. But I guess that it was created to fix these issues. We tested it and it worked :)

[shntnu]

@pamzgt cytominer-database will use the same as your CSV files for the corresponding tables. So if your CSV files are called Per_Cells.csv, the table will be called Per_cells (if you use the latest version, because of this line).

But the fix is simple!

sqlite3 backend.sqlite 'alter table Per_cells rename to Cells'

See https://sqlite.org/lang_altertable.html

[gwaybio]

update: I needed to solve this in #111 - working on it now

[gwaybio]

turns out that the "good first issue" tag here was wrong 😂

Long story short, this is now fully modular (and fully tested) in #111

One key block preventing an easy fix is here

This is now implemented:

from pycytominer.cyto_utils.cells import SingleCells

# Only SQLite supported currently
sc = SingleCells(file_or_conn="sqlite:///EXAMPLE_PATH.sqlite")

# Merge single cell data using the appropriate linking columns
# Linking columns can also be specified for modular compartments
# Different normalization methods also available (see normalize.py)
sc_df = sc.merge_single_cells(single_cell_normalize=False)

# For the default linking cols and for specific modular formatting, run the following
from pycytominer.cyto_utils import get_default_linking_cols
get_default_linking_cols()

# Aggregate single cells
level_3_df = sc.aggregate_profiles()

# Note that above is a faster implementation of the following operation:
from pycytominer import aggregate
level_3_df = aggregate(sc_df)