PEYMA.py

import json
import datasets
import os

_CITATION = """\\
@article{shahshahani2018peyma,
    title={PEYMA: A Tagged Corpus for Persian Named Entities},
    author={Mahsa Sadat Shahshahani and Mahdi Mohseni and Azadeh Shakery and Heshaam Faili},
    year=2018,
    journal={ArXiv},
    volume={abs/1801.09936}
}
"""
_DESCRIPTION = """PEYMA dataset includes 7,145 sentences with a total of 302,530 tokens from which 41,148 tokens are tagged with seven different classes."""

_DATA_PATH = {
  'train': os.path.join('data', 'train.txt'),
  'test': os.path.join('data', 'test.txt'),
  'val': os.path.join('data', 'dev.txt')
}

class PEYMAConfig(datasets.BuilderConfig):
    """BuilderConfig for PEYMA."""
    def __init__(self, **kwargs):
        super(PEYMAConfig, self).__init__(**kwargs)


class PEYMA(datasets.GeneratorBasedBuilder):
    BUILDER_CONFIGS = [
        PEYMAConfig(name="PEYMA", version=datasets.Version("1.0.0"), description="persian ner dataset"),
    ]

    def _info(self):
        return datasets.DatasetInfo(
            # This is the description that will appear on the datasets page.
            description=_DESCRIPTION,
            # datasets.features.FeatureConnectors
            features=datasets.Features(
                {
                "tokens": datasets.Sequence(datasets.Value("string")),
                "tags": datasets.Sequence(
                    datasets.ClassLabel(
                        names=[
                          "O",
                          "B_DAT",
                          "B_LOC",
                          "B_MON",
                          "B_ORG",
                          "B_PCT",
                          "B_PER",
                          "B_TIM",
                          "I_DAT",
                          "I_LOC",
                          "I_MON",
                          "I_ORG",
                          "I_PCT",
                          "I_PER",
                          "I_TIM",
                          ]
                    )
                ),
                }
            ),
            supervised_keys=('tokens', 'tags'),
            # Homepage of the dataset for documentation
            homepage="https://hooshvare.github.io/docs/datasets/ner#peyma",
            citation=_CITATION,
        )
        
    def _split_generators(self, dl_manager):
        """Returns SplitGenerators."""
        return [
                datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": dl_manager.download_and_extract(_DATA_PATH["train"]),
                    "split": "train",
                },),
                datasets.SplitGenerator(
                    name=datasets.Split.TEST,
                    # These kwargs will be passed to _generate_examples
                    gen_kwargs={
                        "filepath": dl_manager.download_and_extract(_DATA_PATH["test"]), 
                        "split": "test"},),
                datasets.SplitGenerator(
                    name=datasets.Split.VALIDATION,
                    # These kwargs will be passed to _generate_examples
                    gen_kwargs={
                        "filepath": dl_manager.download_and_extract(_DATA_PATH["val"]),
                        "split": "validation",
                    },
                ),    
                ]

    def _generate_examples(self, filepath, split):
        with open(filepath, "r", encoding="utf-8") as f:
            id_ = 0
            tokens = []
            ner_labels = []
            for line in f:
                stripped_line = line.strip(" \n")  # strip away whitespaces AND new line characters
                if len(stripped_line) == 0:
                    # If line is empty, it means we reached the end of a sentence.
                    # We can yield the tokens and labels
                    if len(tokens) > 0 and len(ner_labels) > 0:
                        yield id_, {
                            "tokens": tokens,
                            "tags": ner_labels,
                        }
                    else:
                        # Do not yield if tokens or ner_labels is empty
                        # It can be the case if several empty lines are contiguous
                        continue
                    # Then we need to increment the _id and reset the tokens and ner_labels list
                    id_ += 1
                    tokens = []
                    ner_labels = []
                else:
                  try:
                    token, ner_label = line.split("|")  # Retrieve token and label
                    tokens.append(token)
                    ner_labels.append(ner_label)
                  except:
                    continue