simplify

A package that creates simplified likelihoods from full likelihoods. Currently, only one form of simplified likelihoods is implemented, but the idea is to implement additional versions of the simplified likelihoods, such that varying degrees of simplification can be supported.

Installation

Follow good practice and start by creating a virtual environment

python3 -m venv simplify

and then activating it

source simplify/bin/activate

Default install

Install the package with pip

python3 -m pip install simplify-hep[contrib]

Development install

If you want to contribute to simplify, install the development version of the package. Fork the repository, clone the fork, and then install

python3 -m pip install --ignore-installed -U -e .[complete]

Next, setup the git pre-commit hook for Black

pre-commit install

You can run all the tests with

python3 -m pytest

How to run

CLI

Run with e.g.

simplify convert < fullLH.json > simplifiedLH.json

or e.g.

curl http://foo/likelihood.json | simplify convert

where fullLH.json is the full likelihood you want to convert into a simplified likelihood. Simplify is able to read/write from/to stdin/stdout.

In python script

You can also use simplify in a python script, e.g. to create some validation and cross-check plots and tables.

import pyhf
import json

import simplify

pyhf.set_backend(pyhf.tensorlib, "minuit")
spec = json.load(open("likelihood.json", "r"))

ws = pyhf.Workspace(spec) # ws from full LH

# get model and data for each ws we just created
model = ws.model(modifier_settings = {"normsys": {"interpcode": "code4"},"histosys": {"interpcode": "code4p"},})
data = ws.data(model)

# run fit
fit_result = simplify.fitter.fit(ws)

plt = simplify.plot.pulls(
    fit_result,
    "plots/"
)

plt = simplify.plot.correlation_matrix(
    fit_result,
    "plots/",
    pruning_threshold=0.1
)

tables = simplify.plot.yieldsTable(
    ws,
    "plots/",
    fit_result,
)

Dependencies

Naturally relies heavily on pyhf. Part of the code for plotting and validating results is inspired by Alexander Held's cabinetry.