|
"""Bilingual Corpus of Arabic-English Parallel Tweets""" |
|
|
|
|
|
import os |
|
|
|
import pandas as pd |
|
|
|
import datasets |
|
|
|
|
|
_CITATION = """\ |
|
@inproceedings{Mubarak2020bilingualtweets, |
|
title={Constructing a Bilingual Corpus of Parallel Tweets}, |
|
author={Mubarak, Hamdy and Hassan, Sabit and Abdelali, Ahmed}, |
|
booktitle={Proceedings of 13th Workshop on Building and Using Comparable Corpora (BUCC)}, |
|
address={Marseille, France}, |
|
year={2020} |
|
} |
|
""" |
|
|
|
_DESCRIPTION = """\ |
|
Twitter users often post parallel tweets—tweets that contain the same content but are |
|
written in different languages. Parallel tweets can be an important resource for developing |
|
machine translation (MT) systems among other natural language processing (NLP) tasks. This |
|
resource is a result of a generic method for collecting parallel tweets. Using the method, |
|
we compiled a bilingual corpus of English-Arabic parallel tweets and a list of Twitter accounts |
|
who post English-Arabic tweets regularly. Additionally, we annotate a subset of Twitter accounts |
|
with their countries of origin and topic of interest, which provides insights about the population |
|
who post parallel tweets. |
|
""" |
|
|
|
_URL = "https://alt.qcri.org/resources/bilingual_corpus_of_parallel_tweets" |
|
|
|
_DATA_URL = "https://alt.qcri.org/wp-content/uploads/2020/08/Bilingual-Corpus-of-Arabic-English-Parallel-Tweets.zip" |
|
|
|
|
|
class ParallelTweetsConfig(datasets.BuilderConfig): |
|
"""BuilderConfig for Arabic-English Parallel Tweets""" |
|
|
|
def __init__(self, description, data_url, citation, url, **kwrags): |
|
""" |
|
Args: |
|
description: `string`, brief description of the dataset |
|
data_url: `dictionary`, dict with url for each split of data. |
|
citation: `string`, citation for the dataset. |
|
url: `string`, url for information about the dataset. |
|
**kwrags: keyword arguments frowarded to super |
|
""" |
|
super(ParallelTweetsConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwrags) |
|
self.description = description |
|
self.data_url = data_url |
|
self.citation = citation |
|
self.url = url |
|
|
|
|
|
class TweetsArEnParallel(datasets.GeneratorBasedBuilder): |
|
BUILDER_CONFIGS = [ |
|
ParallelTweetsConfig(name=name, description=_DESCRIPTION, data_url=_DATA_URL, citation=_CITATION, url=_URL) |
|
for name in ["parallelTweets", "accountList", "countryTopicAnnotation"] |
|
] |
|
BUILDER_CONFIG_CLASS = ParallelTweetsConfig |
|
|
|
def _info(self): |
|
features = {} |
|
if self.config.name == "parallelTweets": |
|
features["ArabicTweetID"] = datasets.Value("int64") |
|
features["EnglishTweetID"] = datasets.Value("int64") |
|
if self.config.name == "accountList": |
|
features["account"] = datasets.Value("string") |
|
if self.config.name == "countryTopicAnnotation": |
|
features["account"] = datasets.Value("string") |
|
countries = ["QA", "BH", "AE", "OM", "SA", "PL", "JO", "IQ", "Other", "EG", "KW", "SY"] |
|
features["country"] = datasets.features.ClassLabel(names=countries) |
|
topics = [ |
|
"Gov", |
|
"Culture", |
|
"Education", |
|
"Sports", |
|
"Travel", |
|
"Events", |
|
"Business", |
|
"Science", |
|
"Politics", |
|
"Health", |
|
"Governoment", |
|
"Media", |
|
] |
|
features["topic"] = datasets.features.ClassLabel(names=topics) |
|
return datasets.DatasetInfo( |
|
description=_DESCRIPTION, |
|
features=datasets.Features(features), |
|
homepage=self.config.url, |
|
citation=_CITATION, |
|
) |
|
|
|
def _split_generators(self, dl_manager): |
|
dl_dir = dl_manager.download_and_extract(self.config.data_url) |
|
dl_dir = os.path.join(dl_dir, "ArEnParallelTweets") |
|
if self.config.name == "parallelTweets": |
|
return [ |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TEST, |
|
gen_kwargs={ |
|
"datafile": os.path.join(dl_dir, "parallelTweets.csv"), |
|
"split": datasets.Split.TEST, |
|
}, |
|
), |
|
] |
|
|
|
if self.config.name == "accountList": |
|
return [ |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TEST, |
|
gen_kwargs={ |
|
"datafile": os.path.join(dl_dir, "accountList.csv"), |
|
"split": datasets.Split.TEST, |
|
}, |
|
), |
|
] |
|
if self.config.name == "countryTopicAnnotation": |
|
return [ |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TEST, |
|
gen_kwargs={ |
|
"datafile": os.path.join(dl_dir, "countryTopicAnnotation.csv"), |
|
"split": datasets.Split.TEST, |
|
}, |
|
), |
|
] |
|
|
|
def _generate_examples(self, **args): |
|
filename = args["datafile"] |
|
if self.config.name == "parallelTweets": |
|
df = pd.read_csv(filename) |
|
for id_, row in df.iterrows(): |
|
yield id_, {"ArabicTweetID": row["ArabicTweetID"], "EnglishTweetID": row["EnglishTweetID"]} |
|
|
|
if self.config.name == "accountList": |
|
df = pd.read_csv(filename, names=["account"]) |
|
for id_, row in df.iterrows(): |
|
yield id_, { |
|
"account": row["account"], |
|
} |
|
|
|
if self.config.name == "countryTopicAnnotation": |
|
df = pd.read_csv(filename) |
|
for id_, row in df.iterrows(): |
|
yield id_, {"account": row["Account"], "country": row["Country"], "topic": row["Topic"]} |
|
|