jfcrenshaw / photerr Goto Github PK

Add a github action for CI and add contributing docs

The indices of the errors are sequential, starting from 0. This creates problems when adding the errors back into the original DataFrame. For example,

import numpy as np
import pandas as pd
from photerr import LsstErrorModel

x = pd.DataFrame(np.arange(21, 30).reshape(-1, 1).astype(float), columns=["u"])

errM = LsstErrorModel()

errM(x.iloc[2:5], random_state=0)

yields

           u     u_err
2  22.998142  0.077355
3  24.004532       NaN
4  24.947432       NaN
0        NaN  0.014667
1        NaN  0.033837

rather than the expected

           u     u_err
2  22.998142  0.014667
3  24.004532  0.033837
4  24.947432  0.077355

I will fix this so that the errors have the correct indexing.

• sigLim should set the limit that flags non-detections
• sigLim=0 should correspond to only negative fluxes flagged
• ndMode should allow "flag" and "sigLim"
• absFlux as a separate keyword should allow you to take the absolute magnitude of the fluxes. absFlux=True combined with sigLim=0 creates the behavior of the previous ndMode="abs"