The problem seems to have been in GenerateFitShower(), which
assumed all of the 1000s exist when averaging intensities.
I changed the code in two places:
1. It now checks the existence of each CIT1K site individually
instead of checking all of them together.
2. For the purpose of building the fitshower object, uninstalled
CIT1000s are treated the same as ignored CIT1000s, i.e. they
are not included in any of the averaging of areas, time, or
intensities.
Looking at the average intensities for CIT1000, they seem to be down by a factor of 2. This is because two sites are unhealthy during the time interval of this set of showers, but they are not being ignored in the following reconstructions.
Sample showers:
I have set up a test page for the new CTShower code. There are two test libraries, one optimized for chiquita and one optimized for chicos. The test page has access to all the best candidates and a subset of the chiquita showers.
Test pageWhile setting up the test page I noticed that there is a difference between the online version 2.34 and the chiquita code I had been using. The online 2.34 is the same as 2.35 except for the LDF. I had in addition changed the SOTC offsets to be 200 and 40 instead of 200 and 1000.
One problem with the new code seems to be that ignored sites still show up on the map. They do, however, seem to be deleted in the reconstruction.
Shower 3, new code, no IGN'd sites
Shower 3, old code, no IGN'd sites
Shower 3, new code, IGN'd sites
Shower3, old code, IGN'd sites
Shower 4, new code, no IGN'd sites
Shower 4, old code, no IGN'd sites
Shower 4, new code, IGN'd sites
Shower4, old code, IGN'd sites
Another problem seems to be that the angle indicator on the map is opaque. These may both be problems related to switching to ROOT 5.
Both problems are now fixed.
