You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
solarposition.hour_angle should have its own test in test_solarposition. Its output is tested in test_bird but that's not the right place for it.
solarposition.hour_angle is used in test_analytical_zenith and test_analytical_azimuth. That's bad practice. The success of the zenith and azimuth functions should not depend on the hour_angle function.
This situation is making it hard for me to track down an issue. There might be something wrong with solarposition.hour_angle timezone handling, but it's hard to say at the moment.
Versions:
pvlib.__version__: 0.6
Additional context
discovered when addressing pandas upgrade issues in #595
The text was updated successfully, but these errors were encountered:
* add test for hour_angle, vectorize
* closes#598 vectorize to make it more efficient
* closes#597 add test
Signed-off-by: Mark Mikofski <[email protected]>
* BUG: use np.int64 works better for older numpy version than python int
* also converting times to int before subtracting works better for older
pandas versions which were not calculating the timedeltas correctly
* remove comment with missing space after hash, add FIXME that explains
why the expected values are slightly different than the SPA calculator
output
* fix hanging indent
* BUG: replace utcoffset() with reliable, efficient approach ...
* ... suggested by @wholmgren (thx!)
* utcoffset() is unpredictable when used with pandas datetime indices
* only predictable with Python datetime objects or pandas Timestamps
* instead replace tzinfo with None to get naive local times, and
calculate difference from tz-aware times to get timezones
* BUG: combine arithmetic to make calculation more efficient
* also use asarray wrapper at return to ensure consistency
* add comments to explain to future maintainers
* stickler
* remove try-except, doesn't work anyway, wait for pandas >=0.15.0
Problems:
solarposition.hour_angle
should have its own test intest_solarposition
. Its output is tested intest_bird
but that's not the right place for it.solarposition.hour_angle
is used intest_analytical_zenith
andtest_analytical_azimuth
. That's bad practice. The success of the zenith and azimuth functions should not depend on the hour_angle function.This situation is making it hard for me to track down an issue. There might be something wrong with
solarposition.hour_angle
timezone handling, but it's hard to say at the moment.Versions:
pvlib.__version__
: 0.6Additional context
discovered when addressing pandas upgrade issues in #595
The text was updated successfully, but these errors were encountered: