Avoiding phoSim indexing failures in Run 3

[rbiswas4]

@drphilmarshall
Is there a plan to avoid indexing failures due to phosim for run3 ? As a reminder in previous runs, we have found and reported that PhoSim tries to index objects by floats, and indices are messed up due to precision problems when these are large. I do not know if phosim was changed to avoid this problem. I forgot to mention this during the meeting.

A possibility could be checking out our own branch of phosim and modfiying the index to a string before writing it out. We can then tag this and use this tagged version. I will not have time to work on it immediately.

[drphilmarshall]

Hmm. Do you know if phosim has been upgraded to use string or integer indices, @TomGlanzman? Does anyone know whether there is an open ticket on the phosim release bitbucket?

[jchiang87]

This doesn't seem to have been issued there:
https://bitbucket.org/phosim/phosim_release/issues

[drphilmarshall]

Good, thanks Jim. Rahul or Bryce, can you please write a phosim ticket and ask for advice regarding a workaround? When you've made it, please post the link to it here so we can follow along. Thanks!

[rbiswas4]

I thought I saw that Bryce had already posted a link but : https://jira.lsstcorp.org/secure/RapidBoard.jspa?rapidView=12&projectKey=PHOSIM&view=detail&selectedIssue=PHOSIM-27

[rbiswas4]

This issue (specific to run 3) should be closed by #340 . However, I think we should track this general issue.

[rbiswas4]

Closed through #340 for run 3 and kept track of in genral in #383