Skip to content

Kaans toys updates (Don't merge yet!)#105

Open
alikaanguven wants to merge 11 commits intosbi-pdffrom
Kaans-toys-updates
Open

Kaans toys updates (Don't merge yet!)#105
alikaanguven wants to merge 11 commits intosbi-pdffrom
Kaans-toys-updates

Conversation

@alikaanguven
Copy link
Copy Markdown

Don't merge yet!
Lot's of things are suboptimal.
Unnecessary codes.
Cluttered directories.
Messy naming conventions.
Outputs to the user directories instead of scratch etc.
Don't merge.

- add generate_toys.py / generate_toys_v2.py to produce unbinned toys from the N2LL Asimov cache and compute per-toy test statistics for a fixed c1 hypothesis
- add check_CL.py / check_CL_v2.py to evaluate the Asimov test statistic and derive expected CL from the toy distribution
- add user/kaan/README.md with basic usage instructions for toy generation and CL checks
- update common/user.py to configure output locations for Kaan
- add some notebooks for logic checks (checks.ipynb, fake_toys.ipynb, inspect_toy_out.ipynb) for validating toys and N2LL behaviour
- include a sample toy generation log (user/kaan/out.log) for reference
- update README
- update generate_toys_v2.py, but keep only for archival.
- add generate_toys_v3.py with dynamic POI/nuisance CLI parameters
- add plot_toys.py to plot the toy distributions
- add v2 toy generation from target PDF sets via POD-basis projection
- add PDF projection utilities
- update toy fitting and plotting scripts for the new workflow
- replace expected-curve plotting with true/approximated target and toy bands
- refresh README and submission helpers for the new toy pipeline
- remove obsolete generate_toys_v5 script
- add toy generation script for direct PDF reweighting to target LHAPDF sets
- update plotting for PDF bands from toy best-fit results
- add helper to project target PDFs onto the POD basis
- update toy-fit submission to use reweighted toy inputs
- refresh fit-results notebook
@schoef
Copy link
Copy Markdown
Member

schoef commented Mar 25, 2026

  • The Fisher information is in LikelihoodExtensions.py and should be used from there. Don't put it back into Likelihood.py. With the remaining changes to this class we will have to be very careful. Please see how much of it we could put into Extensions.

Otherwise, it is in the user directory, so fine. Should probably just be a "toys" directory.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants