search

openradar / xradar

A tool to work in weather radar data in xarray
https://docs.openradarscience.org/projects/xradar
MIT License
85 stars 17 forks source link

Cf/Radial1 writer #124

Closed syedhamidali closed 9 months ago

syedhamidali commented 10 months ago

Based on the discussions held during the AMS developers meeting, there is a perceived need for a to_cfradial1 function in the Cf/Radial1 format. The decision of its future relevance may depend on Radar Community's compatibility with Cf/Radial1. If necessary, we can consider deprecating it later. I'm eager to collaborate with fellow developers to contribute to this function, not only for its immediate purpose but also to gain experience in contributing to other functions down the line.

kmuehlbauer commented 10 months ago

Should all Cfradial1 versions be supported? Or should we concentrate on a specific version?

mgrover1 commented 10 months ago

I think it is worth supporting both... since we would have a way to validate v1 vs. v2? Open to thought here.

syedhamidali commented 10 months ago

Hello @kmuehlbauer 👋🏻 Which ever version is supported by the majority of the softwares out there, may be we could use the one that is in pyart.io.write

kmuehlbauer commented 10 months ago

Hi @syedhamidali, it might be a good idea to stay in xarray-land also with the transform into CfRadial1.

So as a first proof of concept we could try to reverse the group-extracting we have in https://docs.openradarscience.org/projects/xradar/en/stable/notebooks/CfRadial1_Model_Transformation.html to recreate the source CfRadial1-file from the CfRadial2/FM301-file (current state of affairs).

While doing this we might get insights on how to treat data with different number of range_gates and even different gate spacing.

syedhamidali commented 10 months ago

You are right! I think you've already done most of the work on it in the backends. Can't say anything about treating the data with different numbers of range_gates and even different gate spacing, may be keeping the spacing constant like in RadXConvert: "-const_ngates". Maybe we can add the parameter const_ngates as Bool, and treat the data by default as it is originally.