Notes on the eSASS telecon of 2017-05-15
Preprocessor:
There are two modes: 1) time tags are transmitted for all CCD frames, including those which contain no events and 2) time tags are only available for the non-empty frames. Mode 1) or 2) must be set via the tmsplit XML configuration file (the mode is not inferred from the data).
IK asked whether and if so how the time tags of the empty frames need to be propagated to the output event file (discussion continued after the telecon: IK will create an eventfile extension containing the time tags).
IK suggested that each tmsplit output file (science, HK, star tracker) should have its own (and potentially different) set of GTI. This seems to make sense. The eSASS event files will still only have event GTI, however.
Status of FITS conversion of startracker data: IK can feed the test dataset from IKI (hex dump, converted to binary by GL/IK) through tmsplit. The Radiokomplex header can be attached to the star tracker frames but there are still some uncertainties about the eROSITA frame IDs for the star tracker frames. IK will ask DC for updates.
TEL chain:
MF suggested that the TEL chain should also run for cameras/eROdays which contain no science data. The current approach is to only trigger the TEL chain for cameras/eROdays with science data and a full set of required star tracker and HK data. Changing this would require some thinking on how best to reconfigure the TEL chain (presumably it should terminate after evprep if no science data are present).
mkgti (now renamed to qualgti): The standard header of the output qualgit extension should be copied over from the input GTI extension.
EXP chain:
Changing the exposure maps: see minutes of science requirements meeting.
DET chain:
ermldet: should create additional columns in the output source lists to hold the aperture counts and associated quantities. Details on the apetool/ermldet interface will be discussed offline.
ersensmap: how should the shapelet PSF be averaged for scan mode sensitivity maps? The current approach is photon-based, but ersensmap works independently of eventfiles (AG may have code in apetool for doing attitude based averaging).
Simulation/detection on field scan observations: so far we didn't test detection on field sizes beyond the standard 3.6x3.6 deg all-sky survey fields. Field scan observations can potentially be much larger. While there are no reasons why larger fields should cause problems besides the larger memory requirement, there could be surprises. In principle the field scan fields could be split into standard all sky survey fields, but both in terms of pipeline organization and convenience for the observers analysing them in one piece would be preferred.
SOU chain:
Srctool lightcurves for varichck: SC, TD and HB to meet to have a look (done). The test dataset currently used by TB may have been created with srctool setting (using field GTI instead of source GTI).
eSASS infrastructure:
Logging and error output (WR* routines): the fixed part of the wr* routines output is now much shorter in command line mode, so it should no not be a problem to use the wr* routines for all screen output.
Homogenizing task parameters: next erbox update will contain new parameter names and a task description (already done by Georg). KD will provide versions of his tasks with a proper parameter interface and logging/error output and he will also provide task description (new task versions will then be included in Hamburg tutorial).
Test harness/unit tests: a lot of test harnesses are still missing.
Standard header keywords: the new header keywords DET_CEN and RA_CEN are already included in GL's latest eroio upload.
As the eSASS developers pages (task and caldb status pages, todo lists, etc.) in AWOB has kind of fallen into disuse, HB will move them to the wiki (to a special eSASSdevel area with suitable access rights). TD proposed to use a commercial/open source ticketing system (under investigation).