|
"""Higgs.""" |
|
|
|
from typing import List |
|
|
|
import datasets |
|
|
|
import pandas |
|
|
|
|
|
VERSION = datasets.Version("1.0.0") |
|
_ORIGINAL_FEATURE_NAMES = [ |
|
"is_boson", |
|
"lepton_pT", |
|
"lepton_eta", |
|
"lepton_phi", |
|
"missing_energy_magnitude", |
|
"missing_energy_phi", |
|
"jet1pt", |
|
"jet1eta", |
|
"jet1phi", |
|
"jet1b-tag", |
|
"jet2pt", |
|
"jet2eta", |
|
"jet2phi", |
|
"jet2b-tag", |
|
"jet3pt", |
|
"jet3eta", |
|
"jet3phi", |
|
"jet3b-tag", |
|
"jet4pt", |
|
"jet4eta", |
|
"jet4phi", |
|
"jet4b-tag", |
|
"m_jj", |
|
"m_jjj", |
|
"m_lv", |
|
"m_jlv", |
|
"m_bb", |
|
"m_wbb", |
|
"m_wwbb" |
|
] |
|
|
|
DESCRIPTION = "Higgs dataset from \"Searching for exotic particles in high-energy physics with deep learning\"." |
|
_HOMEPAGE = "https://www.nature.com/articles/ncomms5308/" |
|
_URLS = ("https://www.openml.org/search?type=data&status=active&id=4532") |
|
_CITATION = """ |
|
@article{baldi2014searching, |
|
title={Searching for exotic particles in high-energy physics with deep learning}, |
|
author={Baldi, Pierre and Sadowski, Peter and Whiteson, Daniel}, |
|
journal={Nature communications}, |
|
volume={5}, |
|
number={1}, |
|
pages={4308}, |
|
year={2014}, |
|
publisher={Nature Publishing Group UK London} |
|
}""" |
|
|
|
|
|
urls_per_split = { |
|
"train": "https://gist.githubusercontent.com/msetzu/99114313deb9cc98318d5940fd536b06/raw/7d4cb798cf460aa240b6b92d971009ae6790e3d0/gistfile1.txt", |
|
} |
|
features_types_per_config = { |
|
"higgs": { |
|
"lepton_pT": datasets.Value("float64"), |
|
"lepton_eta": datasets.Value("float64"), |
|
"lepton_phi": datasets.Value("float64"), |
|
"missing_energy_magnitude": datasets.Value("float64"), |
|
"missing_energy_phi": datasets.Value("float64"), |
|
"jet1pt": datasets.Value("float64"), |
|
"jet1eta": datasets.Value("float64"), |
|
"jet1phi": datasets.Value("float64"), |
|
"jet1b-tag": datasets.Value("float64"), |
|
"jet2pt": datasets.Value("float64"), |
|
"jet2eta": datasets.Value("float64"), |
|
"jet2phi": datasets.Value("float64"), |
|
"jet2b-tag": datasets.Value("float64"), |
|
"jet3pt": datasets.Value("float64"), |
|
"jet3eta": datasets.Value("float64"), |
|
"jet3phi": datasets.Value("float64"), |
|
"jet3b-tag": datasets.Value("float64"), |
|
"jet4pt": datasets.Value("float64"), |
|
"jet4eta": datasets.Value("float64"), |
|
"jet4phi": datasets.Value("float64"), |
|
"jet4b-tag": datasets.Value("float64"), |
|
"m_jj": datasets.Value("float64"), |
|
"m_jjj": datasets.Value("float64"), |
|
"m_lv": datasets.Value("float64"), |
|
"m_jlv": datasets.Value("float64"), |
|
"m_bb": datasets.Value("float64"), |
|
"m_wbb": datasets.Value("float64"), |
|
"m_wwbb": datasets.Value("float64"), |
|
"is_boson": datasets.ClassLabel(num_classes=2, names=("no", "yes")) |
|
} |
|
} |
|
features_per_config = {k: datasets.Features(features_types_per_config[k]) for k in features_types_per_config} |
|
|
|
|
|
class HiggsConfig(datasets.BuilderConfig): |
|
def __init__(self, **kwargs): |
|
super(HiggsConfig, self).__init__(version=VERSION, **kwargs) |
|
self.features = features_per_config[kwargs["name"]] |
|
|
|
|
|
class Higgs(datasets.GeneratorBasedBuilder): |
|
|
|
DEFAULT_CONFIG = "higgs" |
|
BUILDER_CONFIGS = [ |
|
HiggsConfig(name="higgs", |
|
description="Higgs boson binary classification.") |
|
] |
|
|
|
|
|
def _info(self): |
|
info = datasets.DatasetInfo(description=DESCRIPTION, citation=_CITATION, homepage=_HOMEPAGE, |
|
features=features_per_config[self.config.name]) |
|
|
|
return info |
|
|
|
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]: |
|
downloads = dl_manager.download_and_extract(urls_per_split) |
|
|
|
return [ |
|
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloads["train"]}) |
|
] |
|
|
|
def _generate_examples(self, filepath: str): |
|
data = pandas.read_csv(filepath) |
|
data = self.preprocess(data, config=self.config.name) |
|
|
|
for row_id, row in data.iterrows(): |
|
data_row = dict(row) |
|
|
|
yield row_id, data_row |
|
|
|
|
|
def preprocess(self, data: pandas.DataFrame, config: str = "higgs") -> pandas.DataFrame: |
|
return data[list(features_types_per_config[config].keys())] |
|
|
