Issue with MVBS calculation for AZFP, missing ping_time dimension

[slonimer]

I am attempting to use echoPype for processing AZFP data. I have extract the data to Sv values, using ep.calibrate.compute_Sv(echodata). However, when I try to calculate MVBS (with ep.preprocess.compute_MVBS or ep.preprocess.compute_MVBS_index_binning), I get the error:
ValueError: Dimensions {'ping_time'} do not exist. Expected one or more of ('frequency', 'range_bin')

I have tried the example for the EK60 data and I am able to calculate MVBS without issue. I have traced the problem to the fact that for the echodata variable, "range", the dimensions are: (frequency, range_bin), whereas for the EK60 data, the dimensions are:
(frequency, ping_time, range_bin).

I have tried to fix this by finding where these dimensions are set for the AZFP, so that I can add the missing ping_time dimension, but I cannot figure this out! Any help would be very much appreciated!

[slonimer]

This doesn't fix the core issue, but I found a work-around where I can add a dimension to the range variable, and then proceed with the MVBS calculations:

ds_Sv["range"] = ds_Sv["range"].expand_dims(axis=1, ping_time = ds_Sv["ping_time"])

[leewujung]

@slonimer Thanks for reporting this! This is unexpected, since an earlier version retained the ping_time dimension. It is possibly due to some reorganization of the calibration functions, since for AZFP a subset of the env params have to be specified by the user, which could come in without any dimensions attached. We are looking into this now!

[leewujung]

FYI -- brief investigation led to documenting the two issues found related to this #453 #454. We'll get this fixed soon.