A tool for chi-squared based kinematic fit hypothesis comparison

[jase231]

Hello! I am trying to integrate the tool described here into the repository.

I am not entirely sure how to integrate my build process into the central Makefile found in the repository root. I have tried patterning my makefile off the environment variables used by some of the other programs' makefiles (e.g. MakeDSelector
/Makefile), but I am not entirely confident they will work. The original makefile, from my standalone repo, can be found here for reference.

Please let me know if anything needs changing!

Best,
Jake Serwe

[aaust]

Hi Jake, thank you for this pull request.
The make file successfully compiles the program. However, it should also copy the executable to $(BMS_OSNAME)/bin, which is in the PATH for our standard environment. You can look at the tree_to_amptools Makefile for inspiration.
As a final step, you should add lines for chisq_hypothesis_comparison into the central Makefile and make_all.sh.

[jase231]

Hi Dr. Austregesilo,
I have made the requested changes. I also modified my Makefile to further reflect the patterns in tree_to_amptools's Makefile (the env target, producing object files in the bin dir, install being the default target).

[aaust]

@jase231 Thanks, that looks good. The Makefile complained about the separators, but I just replaced the spaces with tab characters and it works now. I will try to test it early next week.

[jase231]

Sorry about that! Must've had my IDE misconfigured.

[aaust]

I made a simple example with pi+pi- and K+K- events and successfully suppressed misidentified kaons:

I tested both modes for preserve_combos, but I would prefer to set the default to true as they are needed for proper accidental subtraction. Others may weight in.

And please update the Installation part of the README file.

[jase231]

Should be complete. Let me know if I should be more thorough in the installation instructions; I'm not familiar with the build process the average user at GlueX would be using.

[aaust]

Thanks, that's enough