lookup_linke_turbidity speed improvement, mat to h5 #442
Conversation
* replace LinkeTurbidity matlab file by hdf5 (gzip 4) * update for using LinkeTurbidity hdf5 instead of matlab file in lookup_linke_turbidity function * tests dependencies updates * update requirements / ci for LinkeTurbidity hdf5 file
wholmgren
left a comment
wholmgren
left a comment
There was a problem hiding this comment.
This looks great! Just a few changes needed.
appveyor.yml
Outdated
|
|
||
| # install depenencies | ||
| - cmd: conda create -n test_env --yes --quiet python=%PYTHON_VERSION% pip numpy scipy pandas nose pytest pytz ephem numba siphon -c conda-forge | ||
| - cmd: conda create -n test_env --yes --quiet python=%PYTHON_VERSION% pip numpy scipy tables pandas nose pytest pytz ephem numba siphon -c conda-forge |
There was a problem hiding this comment.
I think you want pytables here and in the env files below.
| Enhancements | ||
| ~~~~~~~~~~~~ | ||
| * | ||
| * Improve clearsky.lookup_linke_turbidity speed (IO) (:issue:`437`) |
There was a problem hiding this comment.
Let's add the .mat to .h5 change to the API changes section.
| try: | ||
| lts = lt_h5_file.root.LinkeTurbidity[latitude_index, longitude_index, :] | ||
| except IndexError: | ||
| raise IndexError('Latitude should be between 90 and -90, ' |
There was a problem hiding this comment.
We should add a test that asserts that this exception is raised under invalid inputs.
Also, I confirmed that the code below does in fact execute the finally block despite a new exception being raised.
try:
raise IndexError
except IndexError:
raise IndexError('blah')
finally:
print('finally')There was a problem hiding this comment.
I actually did write a test for that, but then realized it was already covered.
For the try except finally, I was hoping open_file to come with a context manager, but it does not look like it is the case and fall back to the old school way to make sure the file get closed whatever happen.
There was a problem hiding this comment.
my bad, thanks for reminding me...
|
Looks like the build is passing now. I can squash commits if needed. |
|
Looks finished to me. I plan to merge tomorrow, so now is the time to comment if anyone has objections. |
wholmgren
changed the title
2 participants
git diff upstream/master -u -- "*.py" | flake8 --diffand/or landscape.io linting service.docs/sphinx/source/api.rstfor API changes.docs/sphinx/source/whatsnewfile for all changes.Note:
git diff upstream/master -u -- "*.py" | flake8 --diffis giving mefatal: bad revision 'upstream/master'. I usepylintthough and it didn't complain on additions.