Retrieve energies of triggered events from dl1 parameters

[mikael10j]

• Remove useless code
• Get triggered energies from dl1 params if possible

These energies allow for a computation of the max effective area.

[codecov]

Codecov Report

Patch coverage: 29.25% and project coverage change: +0.03 🎉

Comparison is base (eafbe37) 63.15% compared to head (126f7b7) 63.19%.

❗ Current head 126f7b7 differs from pull request most recent head 31aa52c. Consider uploading reports for the commit 31aa52c to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #188      +/-   ##
==========================================
+ Coverage   63.15%   63.19%   +0.03%     
==========================================
  Files          21       21              
  Lines        2717     2733      +16     
==========================================
+ Hits         1716     1727      +11     
- Misses       1001     1006       +5     

Impacted Files	Coverage Δ
ctaplot/gammaboard/reader.py	0.00% <0.00%> (ø)
ctaplot/plots/style.py	100.00% <ø> (ø)
ctaplot/gammaboard/gammaboard.py	11.81% <2.97%> (-0.10%)	⬇️
ctaplot/ana/ana.py	88.20% <85.71%> (-0.03%)	⬇️
ctaplot/__init__.py	100.00% <100.00%> (ø)
ctaplot/ana/tests/test_ana.py	100.00% <100.00%> (ø)
ctaplot/plots/plots.py	78.46% <100.00%> (ø)
ctaplot/plots/tests/test_plots.py	98.73% <100.00%> (+0.07%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[kosack]

Just a note: you might want to migrate to using ctapipe.io.TableLoader to read the data - it's designed for that purpose and would replace most of the code here. Or at the very least, use ctapipe.io.read_table instead of doing it all manually.

[vuillaut]

Just a note: you might want to migrate to using ctapipe.io.TableLoader to read the data - it's designed for that purpose and would replace most of the code here. Or at the very least, use ctapipe.io.read_table instead of doing it all manually.

Is TableLoader tight to ctapipe data model?
We are mostly using files following lstchain data model at the moment...

[kosack]

Is TableLoader tight to ctapipe data model?
We are mostly using files following lstchain data model at the moment...

Yes, it's basically kept up to date with changes to the ctapipe model (though it's only loosly tied - it doesn't care about the contents of the tables, but the table names do assume ctapipe conventions. The best would be just to swap any write function in lstchain with ctapipe.io.DataWriter, so you get the ctapipe data model.

Alternately, you should a least use ctapipe.io.read_table, which is lower-level, and doesn't assume the ctapipe data model (but does exactly the inverse transform of what ctapipe.io.TableWriter does when writing). I assume LSTChain uses TableWriter? Or is it's IO completely custom?

[vuillaut]

similar to our discussion, this can be simplified using axis

mask = np.unique(dl1_params[['obs_id', 'event_id']], axis=0, return_index=True)
trig_energies = dl1_params['mc_energy'][mask]

[mikael10j]

done