|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
"""TODO: Add a description here.""" |
|
|
|
|
|
import xml.etree.ElementTree as ET |
|
import os |
|
|
|
import datasets |
|
import datasets.features.features |
|
from datasets import ClassLabel |
|
|
|
|
|
_CITATION = """\ |
|
@inproceedings{pontiki-etal-2015-semeval, |
|
title = "{S}em{E}val-2015 Task 12: Aspect Based Sentiment Analysis", |
|
author = "Pontiki, Maria and |
|
Galanis, Dimitris and |
|
Papageorgiou, Haris and |
|
Manandhar, Suresh and |
|
Androutsopoulos, Ion", |
|
booktitle = "Proceedings of the 9th International Workshop on Semantic Evaluation ({S}em{E}val 2015)", |
|
month = jun, |
|
year = "2015", |
|
address = "Denver, Colorado", |
|
publisher = "Association for Computational Linguistics", |
|
url = "https://aclanthology.org/S15-2082", |
|
doi = "10.18653/v1/S15-2082", |
|
pages = "486--495", |
|
} |
|
""" |
|
|
|
_DESCRIPTION = """\ |
|
These are the datasets for Aspect Based Sentiment Analysis (ABSA), Task 12 of SemEval-2015. |
|
""" |
|
|
|
_HOMEPAGE = "https://alt.qcri.org/semeval2015/task12/index.php?id=data-and-tools" |
|
|
|
|
|
_LICENSE = "" |
|
|
|
|
|
|
|
|
|
_URLS = { |
|
"restaurants": {"trial": "absa-2015_restaurants_trial.xml", |
|
"train": "ABSA15_RestaurantsTrain/ABSA-15_Restaurants_Train_Final.xml", |
|
"test": "ABSA15_Restaurants_Test.xml"}, |
|
"laptops": {"trial": "absa-2015_laptops_trial.xml", |
|
"train": "ABSA15_LaptopsTrain/ABSA-15_Laptops_Train_Data.xml", |
|
"test": "ABSA15_Laptops_Test.xml"}, |
|
"hotels": {"test": "ABSA15_Hotels_Test.xml"}, |
|
} |
|
|
|
|
|
class SemEval2015Task12(datasets.GeneratorBasedBuilder): |
|
"""These are the datasets for Aspect Based Sentiment Analysis (ABSA), Task 12 of SemEval-2015.""" |
|
|
|
VERSION = datasets.Version("1.1.0") |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
BUILDER_CONFIGS = [ |
|
datasets.BuilderConfig(name="restaurants", version=VERSION, description="Restaurant reviews"), |
|
datasets.BuilderConfig(name="laptops", version=VERSION, description="Laptop reviews"), |
|
datasets.BuilderConfig(name="hotels", version=VERSION, description="Hotel reviews"), |
|
] |
|
|
|
|
|
|
|
def _info(self): |
|
|
|
categories = { |
|
"restaurants": { |
|
"entities": ["RESTAURANT", "FOOD", "DRINKS", "AMBIENCE", "SERVICE", "LOCATION"], |
|
"attributes": ["GENERAL", "PRICES", "QUALITY", "STYLE_OPTIONS", "MISCELLANEOUS"] |
|
}, |
|
"laptops": { |
|
"entities": ["LAPTOP", "DISPLAY", "KEYBOARD", "MOUSE", "MOTHERBOARD", "CPU", "FANS_COOLING", "PORTS", |
|
"MEMORY", "POWER_SUPPLY", "OPTICAL_DRIVES", "BATTERY", "GRAPHICS", "HARD_DISC", |
|
"MULTIMEDIA_DEVICES", "HARDWARE", "SOFTWARE", "OS", "WARRANTY", "SHIPPING", "SUPPORT", |
|
"COMPANY"], |
|
"attributes": ["GENERAL", "PRICE", "QUALITY", "OPERATION_PERFORMANCE", "USABILITY", "DESIGN_FEATURES", |
|
"PORTABILITY", "CONNECTIVITY", "MISCELLANEOUS"] |
|
}, |
|
"hotels": { |
|
"entities": ["HOTEL", "ROOMS", "FACILITIES", "ROOMS_AMENITIES", "SERVICE", "LOCATION", "FOOD_DRINKS"], |
|
"attributes": ["GENERAL", "PRICES", "COMFORT", "CLEANLINESS", "QUALITY", "DESIGN_FEATURES", |
|
"STYLE_OPTIONS", "MISCELLANEOUS"] |
|
}, |
|
} |
|
polarities = ["positive", "negative", "neutral"] |
|
if self.config.name == "restaurants": |
|
features = datasets.Features( |
|
{ |
|
"reviewId": datasets.Value(dtype="string"), |
|
"sentences": [ |
|
{ |
|
"sentenceId": datasets.Value("string"), |
|
"text": datasets.Value("string"), |
|
"opinions": [ |
|
{ |
|
"target": datasets.Value("string"), |
|
"category": { |
|
"entity": datasets.Value("string"), |
|
"attribute": datasets.Value("string") |
|
}, |
|
"polarity": datasets.Value("string"), |
|
"from": datasets.Value("string"), |
|
"to": datasets.Value("string"), |
|
} |
|
] |
|
} |
|
] |
|
} |
|
) |
|
elif self.config.name == "laptops": |
|
features = datasets.Features( |
|
{ |
|
"reviewId": datasets.Value(dtype="string"), |
|
"sentences": [ |
|
{ |
|
"sentenceId": datasets.Value("string"), |
|
"text": datasets.Value("string"), |
|
"opinions": [ |
|
{ |
|
"category": { |
|
"entity": datasets.Value("string"), |
|
"attribute": datasets.Value("string") |
|
}, |
|
"polarity": datasets.Value("string"), |
|
} |
|
] |
|
} |
|
] |
|
} |
|
) |
|
elif self.config.name == "hotels": |
|
features = datasets.Features( |
|
{ |
|
"reviewId": datasets.Value(dtype="string"), |
|
"sentences": [ |
|
{ |
|
"sentenceId": datasets.Value("string"), |
|
"text": datasets.Value("string"), |
|
"opinions": [ |
|
{ |
|
"target": datasets.Value("string"), |
|
"category": { |
|
"entity": datasets.Value("string"), |
|
"attribute": datasets.Value("string") |
|
}, |
|
"polarity": datasets.Value("string"), |
|
"from": datasets.Value("string"), |
|
"to": datasets.Value("string"), |
|
} |
|
] |
|
} |
|
] |
|
} |
|
) |
|
return datasets.DatasetInfo( |
|
|
|
description=_DESCRIPTION, |
|
|
|
features=features, |
|
|
|
|
|
|
|
|
|
homepage=_HOMEPAGE, |
|
|
|
license=_LICENSE, |
|
|
|
citation=_CITATION, |
|
) |
|
|
|
def _split_generators(self, dl_manager): |
|
|
|
|
|
|
|
|
|
|
|
|
|
urls = _URLS[self.config.name] |
|
data_dir = dl_manager.download_and_extract(urls) |
|
if self.config.name in ["restaurants", "laptops"]: |
|
return [ |
|
datasets.SplitGenerator( |
|
name=datasets.Split("trial"), |
|
|
|
gen_kwargs={ |
|
"filepath": data_dir['trial'], |
|
"split": "trial", |
|
}, |
|
), |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TRAIN, |
|
|
|
gen_kwargs={ |
|
"filepath": data_dir['train'], |
|
"split": "train", |
|
}, |
|
), |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TEST, |
|
|
|
gen_kwargs={ |
|
"filepath": data_dir['test'], |
|
"split": "test" |
|
}, |
|
), |
|
] |
|
elif self.config.name == "hotels": |
|
return [ |
|
datasets.SplitGenerator( |
|
name=datasets.Split.TEST, |
|
|
|
gen_kwargs={ |
|
"filepath": data_dir['test'], |
|
"split": "test" |
|
}, |
|
), |
|
] |
|
|
|
|
|
def _generate_examples(self, filepath, split): |
|
|
|
|
|
tree = ET.parse(filepath) |
|
root = tree.getroot() |
|
for id_, review in enumerate(root.iter("Review")): |
|
reviewId = review.attrib.get("rid") |
|
sentences = [] |
|
for sentence in review.iter("sentence"): |
|
sentence_dict = {} |
|
sentence_dict["sentenceId"] = sentence.get("id") |
|
sentence_dict["text"] = sentence.find("text").text |
|
opinions = [] |
|
for opinion in sentence.iter("Opinion"): |
|
opinion_dict = opinion.attrib |
|
opinion_dict["category"] = dict(zip(["entity", "attribute"], opinion_dict["category"].split("#"))) |
|
opinions.append(opinion_dict) |
|
sentence_dict["opinions"] = opinions |
|
sentences.append(sentence_dict) |
|
yield id_, { |
|
"reviewId": reviewId, |
|
"sentences": sentences |
|
} |
|
|