glass.py

from typing import List
from functools import partial

import datasets

import pandas


VERSION = datasets.Version("1.0.0")


DESCRIPTION = "Glass efficiency dataset from the UCI repository."
_HOMEPAGE = "https://archive-beta.ics.uci.edu/dataset/242/glass+efficiency"
_URLS = ("https://archive-beta.ics.uci.edu/dataset/30/glass+method+choice")
_CITATION = """
@misc{misc_glass_efficiency_242,
  author       = {Tsanas,Athanasios & Xifara,Angeliki},
  title        = {{Glass efficiency}},
  year         = {2012},
  howpublished = {UCI Machine Learning Repository},
  note         = {{DOI}: \\url{10.24432/C51307}}
}"""

# Dataset info
_BASE_FEATURE_NAMES = [
	"refractive_index",
	"sodium",
	"magnesium",
	"aluminum",
	"silicon",
	"potassium",
	"calcium",
	"barium",
	"iron",
	"glass_type",
]
urls_per_split = {
	"train": "https://huggingface.co/datasets/mstz/glass/raw/main/glass.data"
}
features_types_per_config = {
	"glass": {
		"refractive_index": datasets.Value("float64"),
		"sodium": datasets.Value("float64"),
		"magnesium": datasets.Value("float64"),
		"aluminum": datasets.Value("float64"),
		"silicon": datasets.Value("float64"),
		"potassium": datasets.Value("float64"),
		"calcium": datasets.Value("float64"),
		"barium": datasets.Value("int8"),
		"iron": datasets.Value("float64"),
		"glass_type": datasets.ClassLabel(num_classes=7, names=("windows_1", "windows_2",
																"vehicle_windows_1", "vehicle_windows_2",
																"containers", "tableware", "headlamps"))
	},
	"windows": {
		"refractive_index": datasets.Value("float64"),
		"sodium": datasets.Value("float64"),
		"magnesium": datasets.Value("float64"),
		"aluminum": datasets.Value("float64"),
		"silicon": datasets.Value("float64"),
		"potassium": datasets.Value("float64"),
		"calcium": datasets.Value("float64"),
		"barium": datasets.Value("int8"),
		"iron": datasets.Value("float64"),
		"is_windows_glass": datasets.ClassLabel(num_classes=2, names=("no", "yes"))
	},
	"vehicles": {
		"refractive_index": datasets.Value("float64"),
		"sodium": datasets.Value("float64"),
		"magnesium": datasets.Value("float64"),
		"aluminum": datasets.Value("float64"),
		"silicon": datasets.Value("float64"),
		"potassium": datasets.Value("float64"),
		"calcium": datasets.Value("float64"),
		"barium": datasets.Value("int8"),
		"iron": datasets.Value("float64"),
		"is_vehicle_glass": datasets.ClassLabel(num_classes=2, names=("no", "yes"))
	},
	"containers": {
		"refractive_index": datasets.Value("float64"),
		"sodium": datasets.Value("float64"),
		"magnesium": datasets.Value("float64"),
		"aluminum": datasets.Value("float64"),
		"silicon": datasets.Value("float64"),
		"potassium": datasets.Value("float64"),
		"calcium": datasets.Value("float64"),
		"barium": datasets.Value("int8"),
		"iron": datasets.Value("float64"),
		"is_container_glass": datasets.ClassLabel(num_classes=2, names=("no", "yes"))
	},
	"tableware": {
		"refractive_index": datasets.Value("float64"),
		"sodium": datasets.Value("float64"),
		"magnesium": datasets.Value("float64"),
		"aluminum": datasets.Value("float64"),
		"silicon": datasets.Value("float64"),
		"potassium": datasets.Value("float64"),
		"calcium": datasets.Value("float64"),
		"barium": datasets.Value("int8"),
		"iron": datasets.Value("float64"),
		"is_tableware_glass": datasets.ClassLabel(num_classes=2, names=("no", "yes"))
	},
	"headlamps": {
	   "refractive_index": datasets.Value("float64"),
		"sodium": datasets.Value("float64"),
		"magnesium": datasets.Value("float64"),
		"aluminum": datasets.Value("float64"),
		"silicon": datasets.Value("float64"),
		"potassium": datasets.Value("float64"),
		"calcium": datasets.Value("float64"),
		"barium": datasets.Value("int8"),
		"iron": datasets.Value("float64"),
		"is_headlamp_glass": 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 GlassConfig(datasets.BuilderConfig):
	def __init__(self, **kwargs):
		super(GlassConfig, self).__init__(version=VERSION, **kwargs)
		self.features = features_per_config[kwargs["name"]]


class Glass(datasets.GeneratorBasedBuilder):
	# dataset versions
	DEFAULT_CONFIG = "glass"
	BUILDER_CONFIGS = [
		GlassConfig(name="glass", description="Glass dataset."),
		GlassConfig(name="windows", description="Is this windows glass?"),
		GlassConfig(name="vehicles", description="Is this vehicles glass?"),
		GlassConfig(name="containers", description="Is this containers glass?"),
		GlassConfig(name="tableware", description="Is this tableware glass?"),
		GlassConfig(name="headlamps", description="Is this headlamps glass?")
	]


	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, header=None)
		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.columns = _BASE_FEATURE_NAMES

		if self.config.name == "windows":
			data = data.rename(columns={"glass_type": "is_windows_glass"})
			data.loc[:, "is_windows_glass"] = data.is_windows_glass.apply(lambda x: 1 if x == 1 or x == 2 else 0)

		
		elif self.config.name == "vehicles":
			data = data.rename(columns={"glass_type": "is_vehicle_glass"})
			data.loc[:, "is_vehicle_glass"] = data.is_vehicle_glass.apply(lambda x: 1 if x in {3, 4} else 0)
		
		elif self.config.name == "containers":
			data = data.rename(columns={"glass_type": "is_container_glass"})
			data.loc[:, "is_container_glass"] = data.is_container_glass.apply(lambda x: 1 if x == 5 else 0)
		
		elif self.config.name == "tableware":
			data = data.rename(columns={"glass_type": "is_tableware_glass"})
			data.loc[:, "is_tableware_glass"] = data.is_tableware_glass.apply(lambda x: 1 if x == 6 else 0)
		
		elif self.config.name == "headlamps":
			data = data.rename(columns={"glass_type": "is_headlamp_glass"})
			data.loc[:, "is_headlamp_glass"] = data.is_headlamp_glass.apply(lambda x: 1 if x == 7 else 0)
		
		else:
			data.loc[:, "glass_type"] = data.glass_type.apply(lambda x: x - 1)
			print(data.glass_type.unique())

		return data