new RF implementation

[Elisa-Visentin]

Merge of the ctapipe 0.19 and new RF implementation (by Torino team)

[Elisa-Visentin]

I'm trying to solve the issue (seems a problem with scipy version). In the meanwhile, we can check the code (and continue the discussion started in #180

[aleberti]

Thanks for the fix of scipy. Indeed docs could not be built yesterday, but tests were passing... From the docs error it was not clear why it was failing. Can you open a PR just for the scipy fix? so that it can be merged in the master and then I can merge in the PRs I opened in the last days.

[Elisa-Visentin]

Done, #195.
Tests were not failing a couple of days ago, so I think the 'wrong' scipy version is automatically installed in the environment since a few hours 😕

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 77.40%. Comparing base (01f273d) to head (3edc9cb).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #194      +/-   ##
==========================================
+ Coverage   77.23%   77.40%   +0.16%     
==========================================
  Files          21       21              
  Lines        2614     2611       -3     
==========================================
+ Hits         2019     2021       +2     
+ Misses        595      590       -5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[aleberti]

Please merge the master here (note that there is a conflict). In that way you can use the cached test data so that the tests are run (faster).

[jsitarek]

I merged master into this branch, no conflicts were reported (rebasing was not possible, but it did not complain about merging), but something strange is happening because tests fail already at creating environment. I do not see anything strange in the code, and in particular nothing has changed in the setup and environment files

[Elisa-Visentin]

@jsitarek Same problem also in the other PRs. I was just discussing this with Maximilian (ctapipe): also their CIs fail. This seem to be due to an issue with one the dependencies (broken package) Not our fault (hopefully)

[Elisa-Visentin]

Now it should be fine. setuptools 71.xxx has some issues with the metadata of dependencies

[jsitarek]

thanks @Elisa-Visentin
In the meantime I tried this branch for example Prod6 files and at least up to DL2 level it works without any problems

[aleberti]

@Elisa-Visentin @jsitarek I understood this branch was already used for the recent analyses. Anything else to be done here? maybe tests missing?

[Elisa-Visentin]

I was just checking it. It 'inherits' from #180 (just adding new ctapipe), where Julian left some (I found 7) comments/requests. They should be fixed now: @jsitarek could you have a look at the comments you left in the old PR? If they are fixed, we can then merge this one and clone the older one