Dolfyn general updates

[jmcvey3]

Fixes for:

• ADV skipped timesteps
• ADV attributes
• proper setting of max gap in ADV cleaning function
• error in integral length scale calculation
• error in despiking function
• move Reynolds stress and cross spectral density calculations to ADV folder
• prevent inaccurate TKE calculation from ADCP velocity input

I was updating this branch from master, so there may be repeated commits in this list, though that shouldn't matter after resolving conflicts and merging.

[jmcvey3]

Apparently I don't have "write access" to fix these conflicts

[ssolson]

Apparently I don't have "write access" to fix these conflicts

I think that should only occur if you are trying to push changes to the Develop branch. You should be able to resolve these conflicts on your local machine via git pull origin Develop Fix the conflicts git merge git push jmcvey dolfyn

[jmcvey3]

@ssolson Two things: currently netcdf4 is limited by the environment.yml file to <1.5.8, and I just discovered that if installing with conda, (conda install -c conda-forge netcdf4), version 1.5.7 is the latest conda allows.

[jmcvey3]

Having issues with netcdf compression encoding as well in tsdat, and the fundamental problem seems to be related to https://github.com/pydata/xarray/discussions/5709

[jmcvey3]

@ssolson Tests are failing because xarray is searching for a package called bottleneck https://pypi.org/project/Bottleneck/ https://docs.xarray.dev/en/v0.8.0/installing.html. Can we add this package to the requirements list?

[ssolson]

Looks like it is an optional xarray package used to speed up working with NaNs which seems like it would be useful for DOLfYN. I checked out the repo and they seem active and have working tests for all current python versions. I think it should be good to add.

[ssolson]

Hey, James how goes this PR? Do you need any help?

[jmcvey3]

Hi Sterling, this update should be good to go, once I'm sure the tests pass online. Related - is there a different process for updating documentation? I can put a pull request in for that as well.

[jmcvey3]

@ssolson I do need help with fixing the "pip-windows-latest" tests. Should be the last thing.

[ssolson]

@jmcvey3 I am working on a fix.

[jmcvey3]

Hi Adam,

Just left comments on the things you flagged. I removed a number of signal processing functions from velocity.py and tools/ because they're built-in to other packages like scipy (but weren't when dolfyn was originally written).

This pull request will add additional functionality to mhkit because only the basic processing functionality is available right now. I have a few more updates to add to this PR (or a new one) from working with Emily's Tanana River data and Levi's Cook Inlet data. I'm waiting on test data from Emily for those updates, and verification from Levi.

James

[akeeste]

@jmcvey3 Thanks for clarifying and helping track where those functions moved to. That resolves all of my comments. If this PR is otherwise ready to go and more large updates are coming, I think we should merge now and create a new PR for the next updates.

[ssolson]

@jmcvey3 could you merge the latest from Develop into this PR to see if it fixes your pip install CI issues (failing tests)?

[jmcvey3]

Great, sounds good to me Adam