Adding simulation info to DL1 files

[vuillaut]

This PR only adds simulation infos to DL1 files following the structure in v5 discussed in cta-observatory/ctapipe#1059
Somehow fixes #119 .

As the DL1 structure is still being discussed, I think we can move forward with the simu info part to at least have this info available to create IRFs.

The events table structure stays untouched so the rest of the code will still work without any change.

I also modify the merge script to append the simu info from each merged file.

[vuillaut]

This will allow to go further and e.g. avoid requiring simtel files for the sensitivity calculation.
This can be done in another PR after changes from #133 have been merged.

[misabelber]

Hi Thomas! I'm taking a look at the format and dl1 seems to be fine, I could read the mc data without problem. However, I ran lst-recopipe to check how the format is inherited in the dl2 file and I'm finding a weird structure. When I walk through the groups of the dl2 file like this:
f1 = tables.open_file(dl2_file)
for group in f1.walk_groups():
print(group)

I obtain this list:
/ (RootGroup) ''
/events (Group) ''
/simulation (Group) ''
/events/LSTCam (Group) ''

Shouldn't the last one have been erased before writing it to the dl2 file?

[vuillaut]

Hi

Hi Thomas! I'm taking a look at the format and dl1 seems to be fine, I could read the mc data without problem. However, I ran lst-recopipe to check how the format is inherited in the dl2 file and I'm finding a weird structure. When I walk through the groups of the dl2 file like this:
f1 = tables.open_file(dl2_file)
for group in f1.walk_groups():
print(group)

I obtain this list:
/ (RootGroup) ''
/events (Group) ''
/simulation (Group) ''
/events/LSTCam (Group) ''

Shouldn't the last one have been erased before writing it to the dl2 file?

Hi @misabelber,

What do you mean the last one should be erased?
events is just a node with here only the LSTCam node inside as usual.
events/LSTCam contains the DL2 table (same as the DL1 table + the reconstructed fields from the reco dl1 to dl2).

[misabelber]

Ahhh, ok, nevermind then, I was misunderstanding the structure. In that case, no problem!