search

open-gamma-ray-astro / gamma-astro-data-formats

Data formats for gamma-ray astronomy
https://gamma-astro-data-formats.readthedocs.io
Creative Commons Attribution 4.0 International
29 stars 27 forks source link

Explain EVENT_TYPE column or remove #120

Closed maxnoe closed 5 years ago

maxnoe commented 5 years ago

From the specs:

EVENT_TYPE type: bit field (in FITS tform=32X)
        Event quality partition.
cdeil commented 5 years ago

This EVENT_TYPE came from the original event spec by @kosack . Karl, maybe you can comment?

Fermi also has EVENT_TYPE, but I think they use INT , not BITFIELD? See my comment here: https://github.com/open-gamma-ray-astro/gamma-astro-data-formats/issues/119#issuecomment-408022392 Probably we should add a glossary entry explaining what the idea behind event type is (same as Fermi).

@MaxNoe or anyone - If you have time to send PRs for the spec, that would be great. If not I'll try to address these comments soon.

cdeil commented 5 years ago

For IACTs or any data, the idea is that the definition of event class and type are left to the instrument. EVENT_CLASS is provenance, or possibly used when looking up IRFs in a CALDB (not specified yet). EVENT_TYPE is not used by science tools at the moment, but in the future could be in a similar way as Fermi, i.e. to associate with IRFs and have multiple types in a joint analysis.

maxnoe commented 5 years ago

If you have time to send PRs for the spec

I would like to. However, I have no Idea what this field is supposed to contain, this is why I created this issue.

cdeil commented 5 years ago

I tried to address this issue. See https://github.com/open-gamma-ray-astro/gamma-astro-data-formats/pull/116#issuecomment-408611796

I don't think I can explain it better. It's just a few people that are starting to prototype and use EVENT_TYPE, really defining it in the spec and using it like e.g. Fermi-LAT did is future work that will likely take another 1-2 years.

OK?

Closing this issue now. If you have a specific suggestion for a change, or to improve the explanation, please re-open or send a PR or open a new issue.