hepdata4ml - Version 1.1

As expected, the previous version was in need of a very quick tune-up to address some shortcomings with the metadata handling. This release does the following: - fixes some handling of metadata related to pileup inputs; - adds handling of citation metadata, for the various algorithms that are used.

There have been some corresponding updates to the file-checking utility in util/tools/check_file.py, you can now explicitly check the metadata contents of a file by running python util/tools/check_file.py -i <filename> --metaData Or, you can simply print out all the citations in BibTex format as

python util/tools/check_file.py -i <filename> --citations

- Python
Published by janTOffermann 10 months ago

hepdata4ml - Version 1.0

This represents the first big release of HEPData4ML (after Version 0.9 which mainly served to "checkpoint" the code since some datasets on Zenodo had been produced with it)!

As touched upon in the notes in #23 , the code has gone quite an overhaul since the last release -- namely shifting from a tool whereby a user could produce a Monte Carlo simulated dataset of single jets, to full events. With the ability to produce datasets with multiple jets per event -- and even multiple jet collections -- this simple shift (combined with a few other bells and whistles) hopefully expands the usefulness of this package a fair bit.

Given that this refactoring has been a one-person project (and done here and there between other things), I'm sure there are still more things to do (and certainly more things to clean up; I think the repo may still be littered with some unused leftovers of the old code). This will probably necessitate another release in the (near?) future, but I think it makes sense to no longer delay this version as things should be pretty much ready-to-go.

For new users, a good place to start is in the tutorials subdirectory, which includes some scripts to give you a head start on producing Monte Carlo samples; it's pretty barebones at the moment and might not actually explain a whole lot, but the configuration files in there are quite annotated so hopefully this "learning by example" will answer some basic questions on how to actually use the software.

- Python
Published by janTOffermann 10 months ago

hepdata4ml - Version 0.9

This release marks a version of the package compatible with the production of the following datasets: - Top Quark Momentum Reconstruction Dataset [ https://zenodo.org/records/7338117 ] - Top Jet W-Momentum Reconstruction Dataset [ https://zenodo.org/records/8197723 ]

I'm making this a sub-v1 release, as I will soon merge in updates from the devel branch which greatly expand the features and scope of this package -- most notably shifting from "single jet" dataset production to full event datasets. The code is also significantly refactored and cleaned up (hence why I'm doing 0.9 -> 1.0 instead of something like 1.0 -> 2.0).

- Python
Published by janTOffermann 10 months ago