hypo.py

"""Hypo Dataset"""

from typing import List
from functools import partial

import datasets

import pandas


VERSION = datasets.Version("1.0.0")

_ENCODING_DICS = {
	"class": {
		"negative": 0,
		"compensatedhypothyroid": 1,
		"secondaryhypothyroid": 2,
		"primaryhypothyroid": 3
	}
}

DESCRIPTION = "Hypo dataset."
_HOMEPAGE = ""
_URLS = ("")
_CITATION = """"""

# Dataset info
urls_per_split = {
	"train": "https://huggingface.co/datasets/mstz/hypo/resolve/main/hypo.data"
}
features_types_per_config = {
	"hypo": {
		"age": datasets.Value("int64"),
		"sex": datasets.Value("string"),
		"on_thyroxine": datasets.Value("bool"),
		"query_on_thyroxine": datasets.Value("bool"),
		"on_antithyroid_medication": datasets.Value("bool"),
		"sick": datasets.Value("bool"),
		"pregnant": datasets.Value("bool"),
		"thyroid_surgery": datasets.Value("bool"),
		"I131_treatment": datasets.Value("bool"),
		"query_hypothyroid": datasets.Value("bool"),
		"query_hyperthyroid": datasets.Value("bool"),
		"lithium": datasets.Value("bool"),
		"goitre": datasets.Value("bool"),
		"tumor": datasets.Value("bool"),
		"hypopituitary": datasets.Value("bool"),
		"psych": datasets.Value("bool"),
		"TSH_measured": datasets.Value("bool"),
		"TSH": datasets.Value("string"),
		"T3_measured": datasets.Value("bool"),
		"T3": datasets.Value("float64"),
		"TT4_measured": datasets.Value("bool"),
		"TT4": datasets.Value("float64"),
		"T4U_measured": datasets.Value("bool"),
		"T4U": datasets.Value("float64"),
		"FTI_measured": datasets.Value("bool"),
		"FTI": datasets.Value("float64"),
		"TBG_measured": datasets.Value("string"),
		"referral_source": datasets.Value("string"),
		"class": datasets.ClassLabel(num_classes=4,
									 names=("negative", "compensated hypothyroid", "secondary hypothyroid", "primary hypothyroid"))
	},
	"has_hypo": {
		"age": datasets.Value("int64"),
		"sex": datasets.Value("string"),
		"on_thyroxine": datasets.Value("bool"),
		"query_on_thyroxine": datasets.Value("bool"),
		"on_antithyroid_medication": datasets.Value("bool"),
		"sick": datasets.Value("bool"),
		"pregnant": datasets.Value("bool"),
		"thyroid_surgery": datasets.Value("bool"),
		"I131_treatment": datasets.Value("bool"),
		"query_hypothyroid": datasets.Value("bool"),
		"query_hyperthyroid": datasets.Value("bool"),
		"lithium": datasets.Value("bool"),
		"goitre": datasets.Value("bool"),
		"tumor": datasets.Value("bool"),
		"hypopituitary": datasets.Value("bool"),
		"psych": datasets.Value("bool"),
		"TSH_measured": datasets.Value("bool"),
		"TSH": datasets.Value("string"),
		"T3_measured": datasets.Value("bool"),
		"T3": datasets.Value("string"),
		"TT4_measured": datasets.Value("bool"),
		"TT4": datasets.Value("float64"),
		"T4U_measured": datasets.Value("bool"),
		"T4U": datasets.Value("float64"),
		"FTI_measured": datasets.Value("bool"),
		"FTI": datasets.Value("float64"),
		"TBG_measured": datasets.Value("string"),
		"referral_source": datasets.Value("string"),
		"class": datasets.ClassLabel(num_classes=2)
	},
}

features_per_config = {k: datasets.Features(features_types_per_config[k]) for k in features_types_per_config}


class HypoConfig(datasets.BuilderConfig):
	def __init__(self, **kwargs):
		super(HypoConfig, self).__init__(version=VERSION, **kwargs)
		self.features = features_per_config[kwargs["name"]]


class Hypo(datasets.GeneratorBasedBuilder):
	# dataset versions
	DEFAULT_CONFIG = "hypo"
	BUILDER_CONFIGS = [
		HypoConfig(name="hypo", description="Hypo for multiclass classification."),
		HypoConfig(name="has_hypo", description="Hypo for 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)

		for row_id, row in data.iterrows():
			data_row = dict(row)

			yield row_id, data_row

	def preprocess(self, data: pandas.DataFrame) -> pandas.DataFrame:
		data.drop("id", axis="columns", inplace=True)
		data.drop("TBG", axis="columns", inplace=True)

		data = data[data.age != "?"]
		data = data[data.sex != "?"]
		data = data[data.TSH != "?"]

		data.loc[data.T3 == "?", "T3"] = -1
		data.loc[data.TT4 == "?", "TT4"] = -1
		data.loc[data.T4U == "?", "T4U"] = -1
		data.loc[data.FTI == "?", "FTI"] = -1

		data = data.infer_objects()

		for feature in _ENCODING_DICS:
			encoding_function = partial(self.encode, feature)
			data[feature] = data[feature].apply(encoding_function)
		
		if self.config.name == "has_hypo":
			data["class"] = data["class"].apply(lambda x: 0 if x == 0 else 1)
			print("has hypo\n\n\n")
		
		print("classes")
		print(data["class"].unique())

		return data[list(features_types_per_config[self.config.name].keys())]

	def encode(self, feature, value):
		if feature in _ENCODING_DICS:
			return _ENCODING_DICS[feature][value]
		raise ValueError(f"Unknown feature: {feature}")