Update DL3 Tool to to use pyirf IRF interpolation functions

[chaimain]

With the upcoming pyirf PR#150 and the merged PR#141, the DL3 Tool needs to be updated according to Max's suggestion in #696.

As soon as the IO functions in pyirf are merged, I will update the new function interpolate_irf in lstchain/irf/hdu_table.py and with the new MC DL2 productions of different zeniths (same RF model) are ready, the functions can be tested and then I can open the PR.

[codecov]

Codecov Report

Patch coverage: 95.35% and project coverage change: +1.00 🎉

Comparison is base (972d908) 71.76% compared to head (ec65c04) 72.77%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #711      +/-   ##
==========================================
+ Coverage   71.76%   72.77%   +1.00%     
==========================================
  Files         120      122       +2     
  Lines       11348    11849     +501     
==========================================
+ Hits         8144     8623     +479     
- Misses       3204     3226      +22     

Impacted Files	Coverage Δ
lstchain/tools/lstchain_create_irf_files.py	94.97% <86.36%> (-0.63%)	⬇️
lstchain/tools/lstchain_create_dl3_file.py	92.45% <88.70%> (-2.05%)	⬇️
lstchain/high_level/interpolate.py	93.42% <93.42%> (ø)
lstchain/io/io.py	79.19% <95.23%> (+0.49%)	⬆️
lstchain/high_level/__init__.py	100.00% <100.00%> (ø)
lstchain/high_level/hdu_table.py	95.04% <100.00%> (+0.07%)	⬆️
lstchain/high_level/tests/test_hdus.py	100.00% <100.00%> (ø)
lstchain/high_level/tests/test_interpolate.py	100.00% <100.00%> (ø)
lstchain/reco/tests/test_utils.py	100.00% <100.00%> (ø)
lstchain/reco/utils.py	75.52% <100.00%> (+0.52%)	⬆️
... and 2 more

... and 2 files with indirect coverage changes

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

[chaimain]

I have separated the interpolation functions in lstchain/irf/interpolate.py. These functions are:

1. compare_irfs - Compare a given list of IRF files with their metadata, binning and cuts (fixed cuts checked only by checking the assigned metadata). I have to generalize this for energy-dependent cuts as well.
2. load_irf_grid - Compiles a list of column data from the given list of IRFs which are to be used for interpolation
3. interpolate_irf - Computes the interpolation from a given list of IRFs and a dict of target values for interpolation along a particular parameter (zenith, azimuth pointing, etc.)

Currently, only effective area and energy dispersion IRFs are interpolated and written into a final IRF.

The IRF Tool is slightly modified to include necessary metadata values for the interpolation.

The DL3 Tool has the following changes:

1. Instead of providing an input IRF file path, the path where the IRFs are stored (input_irf_path) is to be provided, a glob search pattern (irf_file_pattern) to look up for either a single IRF or multiple for interpolation and the final filename of the interpolated IRF (final_irf_file). This puts the interpolation function as an optional feature.
2. If the interpolation is to be done only on the zenith pointing, a flag interp_only_zd is to be added. By default, I have put zenith and azimuth as the interpolation parameters.
3. There are a few checks on the number of IRFs provided. For interpolation, it has to be at least twice the number of interpolation parameters. The given list of IRFs are also compared first to check if the select metadata, event binning and selection are compatible to be used for interpolation.
4. The target values for interpolation is provided from create_event_list function and then passed into the interpolate_irf function.

For testing these functions, I have temporarily created 2 different IRFs, one with different g/h cut and another with an extrapolated values to check interpolation function.

This PR still needs further cleaning and more improvements.

[gabemery]

First quick pass over the code on my side. I will look more into the complex code later.

[gabemery]

Reviewed interpolate.py

[gabemery]

I now went other the vast majority of the code (not always looking into the details).
Once current items are addressed I think the PR will be ready.

[gabemery]

At this stage I think the PR is ready for merging.