Change way to access monitoring info in dl1 file

[moralejo]

The former approach np.array(f.root[].cols.charge_mean) results in some cases in an attempt to allocate a huge array (instead of just 2x2x1855), and hence a crash. So far I only had this problem at PIC, using lstchain 0.9.13. No idea why...
With current main branch the same code works well at the IT cluster.

Since the new lines work in both cases, I propose to merge the change.

[maxnoe]

Is this by chance the same issue as this here?

cta-observatory/dl1-data-handler#121

[moralejo]

Is this by chance the same issue as this here?

cta-observatory/dl1-data-handler#121

Yes, seems to be the same!

[moralejo]

In my case this is completely deterministic, but I only get it with a v0.9.13 installation at PIC.

[maxnoe]

What is the pytables version there and how was it installed?

[maxnoe]

Since it's when calling np.array on a pytables data structure, it might also depend on the numpy version.

Anyways, the change looks good and is what is actually documented in the pytables docs

[codecov]

Codecov Report

Patch and project coverage have no change.

Comparison is base (d54fdb3) 73.91% compared to head (63d61ab) 73.91%.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1133   +/-   ##
=======================================
  Coverage   73.91%   73.91%           
=======================================
  Files         125      125           
  Lines       12547    12547           
=======================================
  Hits         9274     9274           
  Misses       3273     3273           

Impacted Files	Coverage Δ
...stchain/calib/camera/pixel_threshold_estimation.py	28.26% <0.00%> (ø)

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

[moralejo]

What is the pytables version there and how was it installed?

Name Version Build Channel
pytables 3.8.0 py38h5dedbee_2 conda-forge

Name Version Build Channel
h5py 3.9.0 nompi_py38hc4a6f91_101 conda-forge

With the lstchain installation instructions for installing a release as user. The release was 0.9.13

[maxnoe]

H5py should not be in the play here at all, numpy would be more interesting to know

[moralejo]

H5py should not be in the play here at all, numpy would be more interesting to know

numpy 1.21.6 py38h1d589f8_0 conda-forge

[maxnoe]

Weird... that's the expected old version of numpy together with the current version of pytables...