JBGreisman
commented
1 year ago yep I agree with you that it should at least be clarified in the docstring that merged DataSet objects should be provided in two-column anomalous format.
I will have to think a bit harder about whether one-column anomalous data should be supported.
Not sure if this is an "issue" or not, but
this works:
`mtz="thermolysin_0.mtz" ds =rs.read_mtz(mtz_path+mtz)
ds.compute_dHKL(inplace=True) ds=ds.loc[(ds.dHKL>=custom_range[0]) & (ds.dHKL< custom_range[1]),ds.columns] ds, labels = ds.assign_resolution_bins(nbins) # adds resolution bins
completeness=rs.stats.completeness.compute_completeness(ds.hkl_to_asu(inplace=False,anomalous=True),\ bins=nbins, anomalous=True) completeness.head(nbins)`
but this fails:
`mtz="thermolysin_0.mtz" ds =rs.read_mtz(mtz_path+mtz)
ds.compute_dHKL(inplace=True) ds=ds.loc[(ds.dHKL>=custom_range[0]) & (ds.dHKL< custom_range[1]),ds.columns] ds, labels = ds.assign_resolution_bins(nbins) # adds resolution bins
ds=ds.stack_anomalous() completeness=rs.stats.completeness.compute_completeness(ds.hkl_to_asu(inplace=False,anomalous=True),\ bins=nbins, anomalous=True) completeness.head(nbins)`
The problem seems to be that the assumption is baked in that the user provides unstacked data such that
in_asu, called on line 107 of compute_completenes() (see https://github.com/Hekstra-Lab/reciprocalspaceship/blob/98409cc8552d22d03d6100c6c1dd01cc369e9647/reciprocalspaceship/stats/completeness.py#L107) can use the +ASU, after which the data are stacked.This is probably perfectly fine behavior. The main thing to do would be to add a comment to the doc string that the output should not be stacked.