ctapipe Bi-Weekly Digest — Activity since 2026-06-11

» cta-observatory/ctapipe | cta-observatory/ctapipe

• cta-observatory/ctapipe repo issues [back to top]

  • Pixel-wise aggregation of histograms #577 [calispac] enhancement

    Hi all, I am wondering if there is any plan to have a class/function to create an histogram along a given axis (e.g....

  • Add optimize based on full cut optimization in pyrif.cuts.optimize_cuts #2771 [maxnoe] enhancement

    Please describe the use case that requires this feature. The new pyirf release includes a full grid-search cut...

  • Update DL1 test dataset for machine learning tests #3031 [maxnoe]

    We should update the dl1 test dataset used for the machine learning unit / integration tests. The current datasets...

    • STSpencer: @maxnoe is there anything I can do to help out with this? What exactly do you need? I guess we...

    • maxnoe: @STSpencer thanks for the offer! I think we actually have already the data needed on the grid, I...

    • maxnoe: I will come back to your offer for other test / reference datasets though.

  • Extend `EventSource.is_compatible` to return a reason #3043 [maxnoe] enhancement

    Please describe the use case that requires this feature. In case EventSource unexpectedly cannot find a compatible...

    • kosack: Should this then collect the failure messages of all plugins? It's not obvious which one is...

    • maxnoe: Should this then collect the failure messages of all plugins? Yes, maybe with a mechanism that...

    • maxnoe: The mechanism could be if a message is provided, it is added to the error message.

  • Add `allowed_tels` option to `ApplyModels`-Tool #2978 [Hckjs] enhancement discussion

    If you currently want to do e.g. some benchmark studies on different array configurations from preprocessed DL1 MC's...

    • Hckjs: I guess this is not as straightforward as I initially thought: Some existing DL2 quantities may...

    • maxnoe: I guess this is not as straightforward as I initially thought: A third point: you would also...

  • Proper support for DVRed waveforms #3046 [maxnoe] enhancement

    Please describe the use case that requires this feature. Currently, the waveform array always has shape (n_channels,...

• cta-observatory/ctapipe repo pull_requests [back to top]

  • Make function to compute true disp public in ctapipe.reco #3014 [maxnoe]

    from ctapipe.io import TableLoader from ctapipe.reco.disp import compute_true_disp with TableLoader(path) as...

    • kosack

      • kosack: Is this really the "true" disp if you mix true pointing position and reconstructed shower...

        • kosack: Ok, it's fine

    • kosack

    • STSpencer

  • Move new definition of telescope optics table to OpticsDescription #3027 [maxnoe]

    • kosack

  • Adding a single documentation page with image parameters (features) #3011 [vuillaut]

    This PR introduces a new entry in the user doc "analysis steps" (could be completed later) including a page with all...

    • maxnoe

    • kosack

  • Add one more type annotation to Field overloads #2995 [maxnoe] no-changelog-needed

    While going over the container definitions in #2204, I noticed one missing case where my IDE was complaining about...

    • kosack

  • Allow astropy 8.0 #3006 [maxnoe]

    • kosack

  • Remove fitshistogrram #3032 [maxnoe]

    Closes #1594

    • kosack: edit: responded to wrong PR

    • kosack

  • feat: Implement true_disp calculation in ctapipe-process + test #2951 [koopatroopa787]

    Description This PR implements the calculation and storage of true_disp parameters (norm and sign) in...

    • maxnoe: We decided to approach this issue differently, by exposing the necessary functions publicly, see #3014.

  • AstroQuantity.physical_type from default_value #3012 [LukasBeiske]

    If an AstroQuantity is defined without a physical_type but including the default_value, the physical_type is now...

    • maxnoe: Sorry, one final time please, promise 🤞

    • kosack

    • maxnoe

  • Do not use events outside fov bins for cut optimization #3013 [LukasBeiske]

    Since we do not use fov bins in the cut optimization yet, events outside the fov bins used for the irfs were used in...

  • added HistogramsAggregator #2996 [TjarkMiener]

    closes #577

    • kosack: Looks good overall, just one change request below to store the axis names rather than require...

      • kosack: I think you could drop this axis_names parameter, and just store the axis names in the metadata...

        • TjarkMiener: axis_names actually depends on the input shape of __call__() so the axis_names can not be a...

        • maxnoe: I think @kosack means that you should store the information in the data, not at the class level

        • maxnoe: So that it can be loaded from the container.meta here. Please also call the input container or...

        • kosack: Yes, the idea was to just write the axis names to the output file via the Container metadata,...

        • TjarkMiener: Great, sounds good! Please check d1ba714 if the implementation of this comment is okay. Many...

    • mexanick

    • maxnoe

    • maxnoe: I think this should be configurable or be determined from the input data. But like this, all...

      • TjarkMiener: This was set because #2996 (comment). As I understood it was not working with arbitrary names?

      • maxnoe: It wasn't working because there was a configuration for the name but other places assumed...

      • maxnoe: I think the issue was that I set the name to something else (i.e. pedvar) but the container...

      • TjarkMiener: I think it was the filling of the histogram here which takes the name as argument.

      • TjarkMiener: We could do something like: hist_object.fill(**{self.hist_axis.name: valid_values})

      • TjarkMiener: Yes worked in 3091a4b. Thanks for the comment! @maxnoe

    • kosack: I think you still have a bug: the axis names when I run with the config below are "axis_1",...

    • kosack: Found the bug: do not use pop here, as it modifies the metadata for all rows! ...

      • kosack: Wait, no that might not be the issue. I'll report back shortly sorry.

      • TjarkMiener: 9e7bb50 set the axis names for the PixelStatisticsCalculatorTool to ["channel", "pixel"]. Let me...

      • kosack: I think the way you are the axis_names is more complex than necessary: it could just be a class...

      • TjarkMiener: I think the pop() is needed because of the init of the Hist() below: axes =...

      • maxnoe: But you mutate the container metadata in that operation. You need to make sure this doesn't...

      • kosack: Also: The bug does indeed only happen on the second chunk loaded in the example I had in the...

      • kosack: 9e7bb50 set the axis names for the PixelStatisticsCalculatorTool to ["channel", "pixel"]. Let me...

      • TjarkMiener: hold on for one sec please @kosack

      • TjarkMiener: I implmented the comment to make the axis names as class traits. It is much simpler and now you...

      • maxnoe: but the bug is still here. You .get() from the container meta, which returns the instance and...

      • maxnoe: In 598 you .get(), in 600, you .pop(), mutating the container.

    • kosack: This is the real bug you need to copy it, not modify it in place: axis_kwargs_meta =...

      • kosack: You still need to fix this - it breaks looping through chunks

    • kosack

    • kosack: It's fine to hard-code the axes to pixel/channel in PixelStatisticsCalculator. THey do not need...

    • maxnoe

    • maxnoe

    • kosack: Looks good now. The only thing missing is storing the outliers (which is supported by Hist). ...

    • maxnoe: @TjarkMiener Let's merge this and do improvements in new contributions.

  • Add full cut opt #3035 [Hckjs]

    • Hckjs: sorry, that was unintentional 😅

  • WIP: run CI with astropy 8.0.0rc2 #3034 [maxnoe]

  • bugfix HistoAgg with existing OutlierDetector #3033 [TjarkMiener] bug

    make call() of ODs work with 3D and 4D

    • TjarkMiener: this should be included v0.31.0 @maxnoe . f you run the PixelStatsTool with HistogramAggregator...

    • maxnoe: Suggested change row_dict =...

    • maxnoe: @TjarkMiener Sorry, the indentation was off in the comment. Need to fix locally and also run pre-commit

    • maxnoe

    • kosack

  • Prepare v0.31.0 #3036 [maxnoe] no-changelog-needed

    • kosack

  • Add DL3 writer #2979 [mdebony]

    Add a DL3 writer for the GADF format to catapipe. This PR is sub element of #2727 to simplify the review.

    • maxnoe: @mdebony please rebase against the main branch to get the fixes for the CI

  • Add a StereoDispCombiner #2731 [Hckjs] enhancement algorithm module:reco

    This is basically the StereoCombiner already implemented here in magic-cta-pipe. It combines the mono DISP...

    • maxnoe: This could be a separate issue, but I'm not sure that the code here would be correct for...

      • LukasBeiske: I've seen some sensitivity curves for divergent pointing using this code here and there indeed...

  • DL3 Event List with EventPreprocessor #3004 [kosack]

    This adds a FeatureSet to the EventPreprocessor for transforming DL2 to DL3 observed events. This is a...

    • kosack: @codex review

    • maxnoe: @codex review

  • fix zip for newer pytest version #3041 [TjarkMiener] maintenance

    • maxnoe: I had already opened a merge request for this here: #3039

    • TjarkMiener: Ah okay, sorry for the noise!

  • Fix pytest deprecation warnings #3039 [maxnoe] maintenance no-changelog-needed

    Fixes the current CI failures after the pytest 9.1 release.

    • TjarkMiener

  • add group of star tracker in the hdf5 data format #3040 [TjarkMiener] hdf5 data format

    • mexanick

    • maxnoe: @TjarkMiener Somehow the CI is not starting here, can you push again?

    • maxnoe

    • mexanick

  • add proper sections in hist aggregation tutorial #3042 [TjarkMiener] documentation-only

    • TjarkMiener: See...

    • maxnoe

    • kosack

  • Fix url for doi badge #3037 [maxnoe] no-changelog-needed documentation-only

    • kosack

  • Move decision about time format into data model definition #3038 [maxnoe]

    • kosack

    • maxnoe: @kosack Sorry, missing docstring and changelog

    • maxnoe: @codex review

    • chatgpt-codex-connector: 💡 Codex Review Here are some automated review suggestions for this pull request. Reviewed...

      • chatgpt-codex-connector: Preserve high-resolution default for unannotated time fields For any existing or user-defined...

        • maxnoe: Useful comment, but this is indeed an intentional change, the default should be low.

  • Use newer dl2 file for model training tests #3044 [maxnoe] no-changelog-needed

    Fixes #3031 I'm not totally happy with how these tests are organized, this would very much benefit from a cleanup...

    • maxnoe: This also required me to generate a small simtel prod6 LaPalma file for the application tests,...

  • Add full cut optimization as introduced in pyirf 0.13 #2789 [LukasBeiske]

    This changes the PointSourceSensitivityOptimizer to use the full cut optimization introduced in pyirf 0.13. The...

    • kosack

    • maxnoe: First line of the docstring should be a short title

      • maxnoe: Also include a link to the pyirf docs please:...

    • maxnoe: First line should be a title. I'd include a link to the pyirf docs here:...

    • maxnoe: Two minor comments on docstrings, otherwise looks good, thanks a lot!

    • maxnoe: I resolved those myself

    • maxnoe

    • kosack

  • Update ctapipe-display-dl1 to show bad pixels (nan) in gray #3047 [maxnoe] no-changelog-needed

created at 2026-06-25 09:31 by dinghy 1.4.1.