Deal with Nan timestamps

[maxnoe]

As discussed in #386, adding support for nan values in unix_tai_to_utc

I also renamed the function unix_tai_to_time as it does not actually convert to utc.
It correctly creates an astropy.Time object, that is still in tai but converts to unix utc timestamp when accessing .unix.

For convenience and clarity, I would suggest to compare times to INVALID_TIME as now defined in the module.

[codecov]

Codecov Report

Merging #388 into master will increase coverage by 0.14%.
The diff coverage is 86.36%.

@@            Coverage Diff             @@
##           master     #388      +/-   ##
==========================================
+ Coverage   45.47%   45.62%   +0.14%     
==========================================
  Files          64       64              
  Lines        4723     4736      +13     
==========================================
+ Hits         2148     2161      +13     
  Misses       2575     2575              

Impacted Files	Coverage Δ
lstchain/reco/r0_to_dl1.py	64.18% <25.00%> (ø)
lstchain/reco/tests/test_utils.py	100.00% <100.00%> (ø)
lstchain/reco/utils.py	67.96% <100.00%> (+2.13%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b00cb78...e97b71b. Read the comment docs.

[morcuended]

Seems fine to me. Do we still need to define missing time values as 0 or does this take into account the possibility of encounter nan values?

[maxnoe]

@morcuended it deals with nan like it would be zero.

So your choice now, what you want timestamp to be if it is missing, the conversion functin now supports both

[morcuended]

@morcuended it deals with nan like it would be zero.

So your choice now, what you want timestamp to be if it is missing, the conversion functin now supports both

not strong preference, I would leave it as it is now with nan provided it works

[rlopezcoto]

what about the interpolation, is it dealing with it properly?

[maxnoe]

Can you point me to the interpolation code?

[morcuended]

Can you point me to the interpolation code?

functions linear_imputer and impute_pointing in lstchain/reco/utils.py which interpolate missing alt/az coordinates directly. Thus missing time values should not be a problem.

[morcuended]

May you update the branch with the last merge fixing the muon output filename so I can test it?

[maxnoe]

Thus missing time values should not be a problem.

These are not missing anymore but 1970-01-01, so maybe you need to explicitly filter with time != INVALID_TIME

[rlopezcoto]

@morcuended did you test this? can we move ahead with the merge?

[morcuended]

@morcuended did you test this? can we move ahead with the merge?

Yes, I tested that the script runs well without problems and files are produced normally. However, I did not have the time to look into the time values themselves.