Closed scottstanie closed 1 year ago
Awesome. Any reason not to do a pypi release for this update?
We should cut for version 0.3.0 for this. Would you like to do it? I just added you to the maintainer role of this repo, which should give you the permission to cut releases, let me know if it did not work please.
Here (https://github.com/insarlab/MintPy/wiki/Check-list-for-code-release) is a checklist I use before cutting for a new release, FYI.
hmm I made the release on github, which I thought would trigger the upload to pypi
workflow (I made a new tag when creating a release... I thought that should "push a tag starting with 'v'" and trigger the workflow?). But it appears to have not triggered, so I'll upload it to pypi using twine.
spoke too soon- looks like it's running now 👍 https://github.com/insarlab/PySolid/actions/runs/4733570882
Great, thank you @scottstanie. The conda-forge feedstock should be also triggered automatically within 12 hours.
Update: conda is triggered in about one hour, that PR just got merged.
Returns the arrays to numpy rather than writing to a text file.
To compare, I ran the two examples in the readme for
point
andgrid
, and i saved thefc
variable in a .npy file.point
, it looks like theres about a 1e-7 difference in the E,N,U components. That might be coming from either a float/double conversion, or maybe a loop iteration different (the fortran code was looping one more than we actually wanted for some reason?)grid
, the points agree within ~ 1e-10Also, not super important, but it looks like the
point
example in the readmewhich takes 4 seconds on the main branch takes about 1.5 seconds now