# coding=utf-8 # Copyright 2020 HuggingFace Datasets Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # Lint as: python3 import datasets _DESCRIPTION = """\ A dataset of Legal Documents from German federal court decisions for Named Entity Recognition. The dataset is human-annotated with 19 fine-grained entity classes. The dataset consists of approx. 67,000 sentences and contains 54,000 annotated entities. """ _HOMEPAGE_URL = "https://github.com/elenanereiss/Legal-Entity-Recognition" _CITATION = """\ @misc{https://doi.org/10.48550/arxiv.2003.13016, doi = {10.48550/ARXIV.2003.13016}, url = {https://arxiv.org/abs/2003.13016}, author = {Leitner, Elena and Rehm, Georg and Moreno-Schneider, Julián}, keywords = {Computation and Language (cs.CL), Information Retrieval (cs.IR), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {A Dataset of German Legal Documents for Named Entity Recognition}, publisher = {arXiv}, year = {2020}, copyright = {arXiv.org perpetual, non-exclusive license} } """ _URL = { "train": "https://raw.githubusercontent.com/elenanereiss/Legal-Entity-Recognition/master/data/ler_train.conll", "dev": "https://raw.githubusercontent.com/elenanereiss/Legal-Entity-Recognition/master/data/ler_dev.conll", "test": "https://raw.githubusercontent.com/elenanereiss/Legal-Entity-Recognition/master/data/ler_test.conll", } _VERSION = "1.0.0" class German_LER(datasets.GeneratorBasedBuilder): VERSION = datasets.Version(_VERSION) def _info(self): return datasets.DatasetInfo( description=_DESCRIPTION, features=datasets.Features( { "id": datasets.Value("string"), "tokens": datasets.Sequence(datasets.Value("string")), "ner_tags": datasets.Sequence( datasets.features.ClassLabel( names=[ "B-AN", "B-EUN", "B-GRT", "B-GS", "B-INN", "B-LD", "B-LDS", "B-LIT", "B-MRK", "B-ORG", "B-PER", "B-RR", "B-RS", "B-ST", "B-STR", "B-UN", "B-VO", "B-VS", "B-VT", "I-AN", "I-EUN", "I-GRT", "I-GS", "I-INN", "I-LD", "I-LDS", "I-LIT", "I-MRK", "I-ORG", "I-PER", "I-RR", "I-RS", "I-ST", "I-STR", "I-UN", "I-VO", "I-VS", "I-VT", "O", ] ) ), "ner_coarse_tags": datasets.Sequence( datasets.features.ClassLabel( names=[ "B-LIT", "B-LOC", "B-NRM", "B-ORG", "B-PER", "B-REG", "B-RS", "I-LIT", "I-LOC", "I-NRM", "I-ORG", "I-PER", "I-REG", "I-RS", "O", ] ) ), }, ), supervised_keys=None, homepage=_HOMEPAGE_URL, citation=_CITATION, ) def _split_generators(self, dl_manager): """Returns SplitGenerators.""" # dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs # It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files. # By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive data_dir = dl_manager.download_and_extract(_URL) return [ datasets.SplitGenerator( name=datasets.Split.TRAIN, # These kwargs will be passed to _generate_examples gen_kwargs={"datapath": data_dir["train"], "split": "train"}, ), datasets.SplitGenerator( name=datasets.Split.TEST, # These kwargs will be passed to _generate_examples gen_kwargs={"datapath": data_dir["test"], "split": "test"}, ), datasets.SplitGenerator( name=datasets.Split.VALIDATION, # These kwargs will be passed to _generate_examples gen_kwargs={"datapath": data_dir["dev"], "split": "dev"}, ), ] def _generate_examples(self, datapath, split): sentence_counter = 0 with open(datapath, encoding="utf-8") as f: current_words = [] current_labels = [] current_coarse_labels = [] for row in f: row = row.rstrip() row_split = row.split() if len(row_split) == 2: token, label = row_split current_words.append(token) current_labels.append(label) # generate coarse-grained tags new_label = "" if label == 'O': new_label = label else: bio, fine_tag = label.split("-") if fine_tag in ['PER', 'RR', 'AN']: new_label = bio + '-PER' elif fine_tag in ['LD', 'ST', 'STR', 'LDS']: new_label = bio + '-LOC' elif fine_tag in ['ORG', 'UN', 'INN', 'GRT', 'MRK']: new_label = bio + '-ORG' elif fine_tag in ['GS', 'VO', 'EUN']: new_label = bio + '-NRM' elif fine_tag in ['VS', 'VT']: new_label = bio + '-REG' else: new_label = label current_coarse_labels.append(new_label) else: if not current_words: continue assert len(current_words) == len(current_labels), "word len doesnt match label length" assert len(current_words) == len(current_coarse_labels), "word len doesnt match coarse label length" sentence = ( sentence_counter, { "id": str(sentence_counter), "tokens": current_words, "ner_tags": current_labels, "ner_coarse_tags": current_coarse_labels, }, ) sentence_counter += 1 current_words = [] current_labels = [] current_coarse_labels = [] yield sentence # last sentence if current_words: sentence = ( sentence_counter, { "id": str(sentence_counter), "tokens": current_words, "ner_tags": current_labels, "ner_coarse_tags": current_coarse_labels, }, ) yield sentence