DM-18318: Create initial subset of timeseries features for DIAObject #43
Conversation
morriscb
force-pushed
the
tickets/DM-18318
branch
from
a059578
to
3e5b13d
Compare
Fix up DiaObject schema.
Add totFlux DiaObject columns. Add totFlux columns to afw schema. Fix TotFluxSigma naming. Change totFlux column order. Fix unittests. Fix up linear fitting.
# The first commit's message is: Hook up StentsonJ calculation. # This is the 2nd commit message: Debug StetsonJ calculation.
Fix case mismatch in column names.
Fix flux masking.
morriscb
force-pushed
the
tickets/DM-18318
branch
from
3e5b13d
to
9d0c77d
Compare
data/ppdb-ap-pipe-schema-extra.yaml
Outdated
| type: FLOAT | ||
| nullable: true | ||
| description: Linear best fit slope of the u band fluxes. | ||
| unit: nJy/MJD |
There was a problem hiding this comment.
should be nJy/day (MJD is a time system, not a time unit). Same for the other slopes here.
| @@ -104,6 +104,7 @@ def make_dia_object_schema(filter_names=None): | |||
| doc='Standard error of uPSFluxMean.') | |||
| schema.addField('uPSFluxSigma', type='F', | |||
| doc='Standard deviation of the distribution of uPSFlux.') | |||
|
|
|||
| @@ -120,6 +121,7 @@ def make_dia_object_schema(filter_names=None): | |||
| doc='Standard error of gPSFluxMean.') | |||
| schema.addField('gPSFluxSigma', type='F', | |||
| doc='Standard deviation of the distribution of gPSFlux.') | |||
|
|
|||
| return ans | ||
|
|
||
|
|
||
| def _stentson_J(fluxes, errors): |
There was a problem hiding this comment.
"stentson" -> "stetson" (someone's name); fix throughout
| totFluxErrors = totFluxErrors[noNanMask] | ||
|
|
||
| if len(totFluxes) == 1: | ||
| dia_object_record['%sFPFluxMean' % filter_name] = totFluxes |
There was a problem hiding this comment.
Let's call these totFluxMean and so on so the link to the DIASource quantity is more clear
There was a problem hiding this comment.
That's what I initially had too, however, FPFlux is coming from the current version of the cat schema and the schema in dax_ppdb. I'm trying to reuse those two as much as possible so things don't get even more confused with the column names. I can add totFlux columns to the schema extra files and just leave the FPF names empty but that might be confusing for folks using the current setup. Hopefully once we have a automated way of making yaml schemas from the DPDD soon.
There was a problem hiding this comment.
I just think FPFlux should refer to a different thing, which is means/errors on the diffim forced fluxes.
There was a problem hiding this comment.
Okay, so would you prefer that I add them in as specifically totFlux here and in the schema files?
There was a problem hiding this comment.
Yes, that would be my preference
| """Compute the Stentson mean of the fluxes which down-weights outliers. | ||
|
|
||
| Weighted biased on an error weighted difference scaled by a constant | ||
| (1/``a``) and raised to the power beta. Higher beta's more harshly penalize |
There was a problem hiding this comment.
"beta's" -> "betas" (or "values of beta")
| Returns | ||
| ------- | ||
| stentsonJ : `float` | ||
| StentsonJ statistic for the input fluxes and errors. |
There was a problem hiding this comment.
I'd add a citation note to Stetson 1996
| Returns | ||
| ------- | ||
| wmean : `float` | ||
| Weighted Stentson mean result. |
There was a problem hiding this comment.
I'd add a citation note to Stetson 1996
Add TOTFlux explicitly. Hook up TOTFlux output. Fix linting.
morriscb
force-pushed
the
tickets/DM-18318
branch
from
296d87a
to
89c5465
Compare
No description provided.