pyTMD.compute.tide_currents needing both u_* and v_* rather than only u_* about pytmd

[maucollu]

Well, you're more than welcome, but it is I that should thank you for the quick reaction!
I suppose the results are in meter per seconds, right?

[maucollu]

Ahhh, that's why...I was getting some pretty big numbers if they were in m/s...thanks a lot.

[tsutterley]

This is all in the recently released v2.1.1. Thanks for your help on this!
-ts

[tsutterley]

I believe that is for backwards compatibility for when TPXO9-atlas models had served the u and v currents in separate files (see #258). I only have access to the binary (OTIS) file type, so I don't know how the netCDF files are currently provided. I can edit the io.model class for TPXO9-atlas-v5 if each netCDF4 file provides both the zonal and meridional currents.

In the compute.tide_currents function, the binary OTIS file type only reads from the u part of the dictionary. So in the mean time, using the binary files should (hopefully) work fine.

[maucollu]

Many thanks, @tsutterley , for helping with this.
Unfortunately, I changed the setting to ATLAS_FORMAT='OTIS'. and it is still looking for the "v_*" types as well, more precisely in:
pyTMD\io\model.py:842 in current

I was trying to identify, in the code, where it actually reads the two components of the velocity, but got a bit lost ;).

Many thanks for the help with this.

Kind Regards

Maurizio

[tsutterley]

you're right that is a problem. It's in the pathfinder function within the io.model class. There's a file check to make sure everything exists. I need to fix this.

[tsutterley]

I'm working on this now. Can you verify for me that the TPXO9-atlas-v5 netCDF4 files have both components?

[maucollu]

I checked, and it seems so.
I quickly did a data = ncinfo('u_2n2_tpxo9_atlas_30_v5.nc') in MATLAB, and in data.Variables I can see:

• uRe
• uIm
• vRe
• vIm
  so I suppose these are the real and imaginary parts of the u and v velocities.
  I'm checking tha tthe same can be seen in python

[maucollu]

Yes, it seems so:
<class 'netCDF4._netCDF4.Dataset'> root group (NETCDF3_CLASSIC data model, file format NETCDF3): type: OTIS transport file title: TPXO9-ATLAS-v5 1/30 x 1/30 tidal SN and WE transports file dimensions(sizes): nx(10800), ny(5401), nct(4) variables(dimensions): |S1 con(nct), float64 lon_u(nx), float64 lat_u(ny), float64 lon_v(nx), float64 lat_v(ny), int32 uRe(nx, ny), int32 uIm(nx, ny), int32 vRe(nx, ny), int32 vIm(nx, ny)

[tsutterley]

Perfect thanks! I'll get this in as part of a PR I started earlier and will put out a release today. Thanks for your help on this. I don't really work with the current files.
And, yes. Those are the real and imaginary parts of the current constituents.

[tsutterley]

should be in cm/s

[maucollu]

Hi, apologies for the "noob" question - but would be 2.1.1 available through the install mode "conda install -c conda-forge pytmd", or do I need to do the git and then python setup.py install? I tried both, but i do not seem to be able to use the new version....

For context, this is what I tried (on top of the git clone + setup):
`conda update pyTMD
Channels:
Note: you may need to restart the kernel to use updated packages.

• conda-forge
• scikit-learn
• bioconda
• nrel
• defaults
  Platform: win-64
  Collecting package metadata (repodata.json): ...working... done
  Solving environment: ...working... done

All requested packages already installed.

conda install -c conda-forge pytmd=2.1.1
Channels:

• conda-forge
• scikit-learn
• bioconda
• nrel
• defaults
  Platform: win-64
  Collecting package metadata (repodata.json): ...working... done
  Solving environment: ...working... failed

Note: you may need to restart the kernel to use updated packages.

PackagesNotFoundError: The following packages are not available from current channels:

• pytmd=2.1.1*

Current channels:

• https://conda.anaconda.org/conda-forge
• https://conda.anaconda.org/scikit-learn
• https://conda.anaconda.org/bioconda
• https://conda.anaconda.org/nrel
• defaults

To search for alternate channels that may provide the conda package you're
looking for, navigate to

https://anaconda.org

and use the search bar at the top of the page.`

[tsutterley]

I think you should be able to do either (depending on your current installation):
conda update pyTMD
or
pip install --upgrade pyTMD
as both of the providers (conda-forge and pypi) are up-to-date with v2.1.1.

[maucollu]

Which version of python is necessary?

[tsutterley]

We test against Python 3.11 and I currently run Python 3.12. But on my old machine I was running an earlier version of python3. I think I have included the imports necessary for earlier versions to work (from 3.6+).

[maucollu]

It works! Many thanks.