The current estimates of varg and vark assume that the fields are noise dominated with a constant variance (e.g. "shape noise" for shear fields), so one can get a good estimate of the appropriate variance empirically from calculating <|g|^2> or <(k-k_mean)^2>.
@AlexGKim has a use case where that might not be the case, but he knows the actual variance in the measurements. So this issue is a feature request to allow the user to set the variance explicitly in a Catalog, rather than let TreeCorr calculate it internally.
Two possible APIs to consider:
Set it on construction: cat = Catalog(..., vark=vark)
The current estimates of varg and vark assume that the fields are noise dominated with a constant variance (e.g. "shape noise" for shear fields), so one can get a good estimate of the appropriate variance empirically from calculating <|g|^2> or <(k-k_mean)^2>.
@AlexGKim has a use case where that might not be the case, but he knows the actual variance in the measurements. So this issue is a feature request to allow the user to set the variance explicitly in a Catalog, rather than let TreeCorr calculate it internally.
Two possible APIs to consider:
cat = Catalog(..., vark=vark)cat.vark = varkOr maybe just allow both would also make sense.