Change underlying computation in `compute_biomass` to xarray operations

[leewujung]

Currently matrix_multiply_grouped_table called under compute_biomass has ops cross between using pandas dataframes and xarray dataset/dataarray. If all inputs uses xr.dataarray as containers, the dimensions would align automatically and this special function is not needed.

[brandynlucca]

As discussed, none of the inputs at xarray.DataArray or xarray.Dataset objects, so no conversion occurs. Consequently, this can be closed.

[leewujung]

Isn't matrix_multiply_grouped_table using dataxarray as input?

echopop/echopop/workflows/nwfsc_feat/biology.py

Lines 287 to 293 in 1b9f3f0

	def matrix_multiply_grouped_table(
	dataset: pd.DataFrame,
	table: xr.DataArray,
	variable: str,
	output_variable: str,
	group: Optional[str] = None,
	) -> None:

[brandynlucca]

Ah yes, I misspoke about "none of the inputs" when I meant specifically the input transect dataset. So yes, the argument table is an xarray.DataArray, but it isn't converted into a pandas.DataFrame within the function, and the dataset argument here corresponds to the transect data which is currently a pandas.DataFrame in the main workflow. It seems that I misunderstood what the issue is/was when we discussed this since I thought we had resolved the confusion around this. The lack of conversion doesn't really address the "ops cross" that you brought up either way, so I think I was overlooking that specific point since I had "get rid of all xarray -> pandas conversions" front-and-center in my mind. I've flagged this issue with the meeting_discussion label I just created to bring up next time we meet to discuss the appropriate remedy here.

[brandynlucca]

I've also re-added the high_priority label.

[brandynlucca]

Based on today's discussion:

• Clarify operation in docstring
• Rename variables to avoid confusion between DataFrame and DataArray/Dataset objects