tic_tac_toe.py

"""TicTacToe"""

from typing import List

import datasets

import pandas


VERSION = datasets.Version("1.0.0")
_BASE_FEATURE_NAMES = [
   "top_left_square",
   "top_middle_square",
   "top_right_square",
   "middle_left_square",
   "middle_middle_square",
   "middle_right_square",
   "bottom_left_square",
   "bottom_middle_square",
   "bottom_right_square",
   "x_wins"
]

DESCRIPTION = "TicTacToe dataset from the UCI ML repository."
_HOMEPAGE = "https://archive.ics.uci.edu/ml/datasets/TicTacToe"
_URLS = ("https://archive.ics.uci.edu/ml/datasets/TicTacToe")
_CITATION = """
@misc{misc_tic-tac-toe_endgame_101,
  author       = {Aha,David},
  title        = {{Tic-Tac-Toe Endgame}},
  year         = {1991},
  howpublished = {UCI Machine Learning Repository},
  note         = {{DOI}: \\url{10.24432/C5688J}}
}"""

# Dataset info
urls_per_split = {
	"train": "https://huggingface.co/datasets/mstz/tic_tac_toe/raw/main/tic-tac-toe.data"
}
features_types_per_config = {
	"tic_tac_toe": {
		"top_left_square": datasets.Value("string"),
		"top_middle_square": datasets.Value("string"),
		"top_right_square": datasets.Value("string"),
		"middle_left_square": datasets.Value("string"),
		"middle_middle_square": datasets.Value("string"),
		"middle_right_square": datasets.Value("string"),
		"bottom_left_square": datasets.Value("string"),
		"bottom_middle_square": datasets.Value("string"),
		"bottom_right_square": datasets.Value("string"),
		"x_wins": 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 TicTacToeConfig(datasets.BuilderConfig):
	def __init__(self, **kwargs):
		super(TicTacToeConfig, self).__init__(version=VERSION, **kwargs)
		self.features = features_per_config[kwargs["name"]]


class TicTacToe(datasets.GeneratorBasedBuilder):
	# dataset versions
	DEFAULT_CONFIG = "tic_tac_toe"
	BUILDER_CONFIGS = [
		TicTacToeConfig(name="tic_tac_toe",
					description="TicTacToe 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, header=None)
		data.columns = _BASE_FEATURE_NAMES
		data.loc[:, "x_wins"] = data.x_wins.apply(lambda x: 1 if x == "positive" else 0)

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

			yield row_id, data_row