Changes between Version 22 and Version 23 of ThingsToBeDoneBeforeSas40

Timestamp:

Sep 23, 2015, 6:08:39 PM (11 years ago)

Author:

eugene

Comment:

--

ThingsToBeDoneBeforeSas40

@@ -99 +99 @@
  * Nigel: there is a population with halflightrad < 0.1. This is not a star/galaxy issue (using SDSS types, everything with rad<~2 is a star). Colours show density of points.  '''asana'''
    * HAF/KCC: as listed in Schema 21 the halflight radius is not treated consistently. There is momenRH (labeld as half light radius, sometimes with DROP), in Detection, diffDetection, ForcedWarp, and Stack. This comes from momentRH in the cmf files. We also have a true halflightrad in stack, which is filled from the cmf (HALF_LIGHT_RADIUS. We need to check that this is present in all types of cmf files and use HALF_LIGHT_RADIUS where it was measured and drop momentRH everywhere.
- * Nigel: psfMajor/MinorFWHM are in pixels but the schema says arcsec. 
+ * Nigel: psfMajor/MinorFWHM are in pixels but the schema says arcsec. '''asana'''
    * HAF: ipptopsps grabs it directly from dvo (no conversions, i see min/majorFWHM is in stack, detection, diff, forcedwarp) - is it wrong for everything or only certain tables? If only certain tables, EAM needs to apply something in dvo. If it's wrong for all, HAF needs to apply correction in ipptopsps. (see other examples, in ipptopsps)
- * Roberto: In ForcedMeanObject all g,r,i,r,y-Gamma values are still -999. 
- * apflux and apfluxerr still follow multiple, noise-free sequences in all filters. The values of apfluxstd are strange, why can they have the value of -999 when apfluxerr is 'normal'? This happens when nTotal>1: for example, obID=ForcedMeanObject in the gband has gnTotal=4, gApFlux=8.19722E-06, gApFluxErr=3.59973E-07 and gApFluxStd=-999.
+ * Roberto: In ForcedMeanObject all g,r,i,r,y-Gamma values are still -999. '''asana, done'''
+    * EAM: these are not valid and should be dropped
+ * apflux and apfluxerr still follow multiple, noise-free sequences in all filters.  '''asana'''
+ * The values of apfluxstd are strange, why can they have the value of -999 when apfluxerr is 'normal'? '''asana'''
+   * This happens when nTotal>1: for example, obID=ForcedMeanObject in the gband has gnTotal=4, gApFlux=8.19722E-06, gApFluxErr=3.59973E-07 and gApFluxStd=-999.
    *  HAF: these come directly from the cmf, no other translations done - EAM? EAM will look at
- * Roberto: in StackModelFitDeV the DeVCf are all -999 in all filters.
- * Roberto: in StackModelFitExp ExpCf are all -999. There are still 'sensible' fits (example: objID=101703338181088339, gband) where RAexp and DECexp are -999
- * Roberto: in StackModelFitSer ExpCf are all -999. nuerr=0.1 always
+   * EAM: I investigated these: the values are correctly inherited from the cmf.  the trends are correct except that the apFluxErr uses sqrt(image) instead of sqrt(variance).  
+ * Roberto: in StackModelFitDeV the DeVCf are all -999 in all filters.  '''asana'''
+ * Roberto: in StackModelFitExp ExpCf are all -999. There are still 'sensible' fits (example: objID=101703338181088339, gband) where RAexp and DECexp are -999  '''asana'''
+ * Roberto: in StackModelFitSer ExpCf are all -999. nuerr=0.1 always  '''asana'''
    * HAF CF is not loaded for Dev/Exp/Ser in ipptopsps, need to investigte if in cmf. EAM : not measured, drop it
- * Roberto: in StackModelFitExtra logC and logA are -999
+ * Roberto: in StackModelFitExtra logC and logA are -999  '''asana'''
    * HAF logC I couldn't find a column for in the CMF, logA is loaded by ipptopsps so must be null in the smf file. EAM: not measured, drop them
- * Roberto: in StackPetrosian many petRadiusErr are -999, although all the other errors are not (example: objID=101663337788271850, gband)
+ * Roberto: in StackPetrosian many petRadiusErr are -999, although all the other errors are not (example: objID=101663337788271850, gband)  '''asana'''
    * HAF PetRadErr is loaded in ipptopsps, might be null in cmf (EAM?) EAM to investigate
- *  Roberto: in ForcedWarpMeasurement psfLikelihood is always -999
+ *  Roberto: in ForcedWarpMeasurement psfLikelihood is always -999  '''asana'''
    * HAF: loaded in ipptopsps: might not be set in CMF (EAM?) EAM: do we have a chi sq? calculated in ipptopsps from a chi sq (HAF to investigate)
- *  Roberto: in Detection the apflux and apfluxerr still follow multiple, noise-free sequences, errors on kron fluxes don't
+ *  Roberto: in Detection the apflux and apfluxerr still follow multiple, noise-free sequences, errors on kron fluxes don't  '''asana'''
    * HAF: EAM investigate?
- *  Roberto: in DiffDetObject cx, cy,cz, l, b, lambda, beta are -999, raErr and DecErr are 0
+ *  Roberto: in DiffDetObject cx, cy,cz, l, b, lambda, beta are -999, raErr and DecErr are 0 '''asana'''
    * HAF: cx/cz/cy/l/b/lambda/beta should be calculated by PSPS.
    * HAF raErr/decErr sounds like dvo calibration problem (bug EAM, ipptopsps tries to load) HAF: need to remove ra/dec err does not make sense for diff