File size: 5,190 Bytes
5d592f8 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 |
"""Europarl Monolingual Dataset."""
from __future__ import absolute_import, division, print_function
from dataclasses import dataclass
import json
import os
import re
import datasets
_CITATION = """\
@inproceedings{koehn2005europarl,
title={Europarl: A parallel corpus for statistical machine translation},
author={Koehn, Philipp},
booktitle={MT summit},
volume={5},
pages={79--86},
year={2005},
organization={Citeseer}
}
"""
_DESCRIPTION = """\
Europarl Monolingual Dataset.
The Europarl parallel corpus is extracted from the proceedings of the
European Parliament (from 2000 to 2011). It includes versions in 21 European
languages: Romanic (French, Italian, Spanish, Portuguese, Romanian),
Germanic (English, Dutch, German, Danish, Swedish), Slavik (Bulgarian,
Czech, Polish, Slovak, Slovene), Finni-Ugric (Finnish, Hungarian, Estonian),
Baltic (Latvian, Lithuanian), and Greek.
Upstream url: https://www.statmt.org/europarl/
"""
_HOMEPAGE = "https://www.statmt.org/europarl/"
_AVAILABLE_LANGUAGES = [
("bg", "Bulgarian"), ("cs", "Czech"), ("da", "Danish"), ("de", "German"),
("el", "Greek"), ("en", "English"), ("es", "Spanish"), ("et", "Estonian"),
("fi", "Finnish"), ("fr", "French"), ("hu", "Hungarian"), ("it", "Italian"),
("lt", "Lithuanian"), ("lv", "Latvian"), ("nl", "Dutch"), ("pl", "Polish"),
("pt", "Portuguese"), ("ro", "Romanian"), ("sk", "Slovak"),
("sl", "Slovene"), ("sv", "Swedish")
]
@dataclass
class EuroparlMonoConfig(datasets.BuilderConfig):
"""BuilderConfig for Europarl Monolingual."""
language: str = None
class EuroparlMonoDataset(datasets.GeneratorBasedBuilder):
"""Europarl Monolingual Dataset."""
_TRAIN_FILE = "train.jsonl"
_VAL_FILE = "val.jsonl"
_TEST_FILE = "test.jsonl"
BUILDER_CONFIGS = [
EuroparlMonoConfig(
name="europarl-%s" % langISO2,
language=langISO2,
version=datasets.Version("7.0.0"),
description="Europarl %s Dataset." % langEnglish,
) for langISO2, langEnglish in _AVAILABLE_LANGUAGES
]
def _info(self):
return datasets.DatasetInfo(
description=_DESCRIPTION,
features=datasets.Features(
{
"paragraph": datasets.Value("string"),
"date": datasets.Value("string"),
"chapter": datasets.Value("string"),
"speaker": datasets.Value("int16"),
"speaker_name": datasets.Value("string")
}
),
supervised_keys=None,
homepage=_HOMEPAGE,
citation=_CITATION,
)
def _split_generators(self, dl_manager):
arch_path = dl_manager.download_and_extract(os.path.join(_HOMEPAGE, "v%d" % self.config.version.major, "europarl.tgz"))
path_dir = os.path.join(arch_path, os.path.join("txt", self.config.language))
paths = [os.path.join(path_dir, file) for file in os.listdir(path_dir)]
return [
datasets.SplitGenerator(
name=datasets.Split.TRAIN, gen_kwargs={"filepath": paths}
)
]
def _generate_examples(self, filepath):
"""Generate Europarl Monolingual examples."""
id_ = 0
for file in filepath:
with open(file, "r") as f:
current_date = "-".join(os.path.basename(file).strip(".txt").split("-")[-4:-1])
current_paragraph = ""
current_paragraph_id = 0
current_chapter = 0
current_speaker_id = 0
current_speaker_name = ""
for line in f:
if not line.startswith("<"):
current_paragraph += line.strip("\n")
else:
if len(current_paragraph) != 0:
yield id_, {
"paragraph": current_paragraph,
"date": current_date,
"chapter": current_chapter,
"speaker": current_speaker_id,
"speaker_name": current_speaker_name,
}
current_paragraph = ""
id_ += 1
if line.startswith("<P>"):
current_paragraph_id += 1
if line.startswith("<CHAPTER"):
current_chapter = line.lstrip("<CHAPTER ID=").rstrip(">\n").strip("\"")
if line.startswith("<SPEAKER"):
current_speaker_id = "0"
current_speaker_name = ""
grps = re.findall("([^<\s]+)=([^\s\">]+|\"[^\">]+\")", line)
for attr, value in grps:
value = value.strip("\"")
if attr == "ID":
current_speaker_id = int(value)
elif attr == "NAME":
current_speaker_name = value
|