# coding=utf-8 # Copyright 2022 The HuggingFace Datasets Authors and Simon Ott, github: nomisto # # 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. """ The LLL05 challenge task is to learn rules to extract protein/gene interactions from biology abstracts from the Medline bibliography database. The goal of the challenge is to test the ability of the participating IE systems to identify the interactions and the gene/proteins that interact. The participants will test their IE patterns on a test set with the aim of extracting the correct agent and target.The challenge focuses on information extraction of gene interactions in Bacillus subtilis. Extracting gene interaction is the most popular event IE task in biology. Bacillus subtilis (Bs) is a model bacterium and many papers have been published on direct gene interactions involved in sporulation. The gene interactions are generally mentioned in the abstract and the full text of the paper is not needed. Extracting gene interaction means, extracting the agent (proteins) and the target (genes) of all couples of genic interactions from sentences. """ # NOTE: # word stop offsets are increased by one to be consistent with python slicing. # test set does not include entity relation information import itertools as it from typing import List import datasets from .bigbiohub import kb_features from .bigbiohub import BigBioConfig from .bigbiohub import Tasks from .bigbiohub import BigBioValues _LANGUAGES = ['English'] _PUBMED = True _LOCAL = False _CITATION = """\ @article{article, author = {Nédellec, C.}, year = {2005}, month = {01}, pages = {}, title = {Learning Language in Logic - Genic Interaction Extraction Challenge}, journal = {Proceedings of the Learning Language in Logic 2005 Workshop at the \ International Conference on Machine Learning} } """ _DATASETNAME = "lll" _DISPLAYNAME = "LLL05" _DESCRIPTION = """\ The LLL05 challenge task is to learn rules to extract protein/gene interactions from biology abstracts from the Medline bibliography database. The goal of the challenge is to test the ability of the participating IE systems to identify the interactions and the gene/proteins that interact. The participants will test their IE patterns on a test set with the aim of extracting the correct agent and target.The challenge focuses on information extraction of gene interactions in Bacillus subtilis. Extracting gene interaction is the most popular event IE task in biology. Bacillus subtilis (Bs) is a model bacterium and many papers have been published on direct gene interactions involved in sporulation. The gene interactions are generally mentioned in the abstract and the full text of the paper is not needed. Extracting gene interaction means, extracting the agent (proteins) and the target (genes) of all couples of genic interactions from sentences. """ _HOMEPAGE = "http://genome.jouy.inra.fr/texte/LLLchallenge" _LICENSE = 'License information unavailable' _URLS = { _DATASETNAME: [ "http://genome.jouy.inra.fr/texte/LLLchallenge/data/LLLChalenge05/data/train/task2/genic_interaction_linguistic_data.txt", # noqa "http://genome.jouy.inra.fr/texte/LLLchallenge/data/LLLChalenge05/data/train/task2/genic_interaction_linguistic_data_coref.txt", # noqa "http://genome.jouy.inra.fr/texte/LLLchallenge/data/LLLChalenge05/data/test/task2/enriched_test_data.txt", # noqa ] } _SUPPORTED_TASKS = [Tasks.RELATION_EXTRACTION] _SOURCE_VERSION = "1.0.0" _BIGBIO_VERSION = "1.0.0" class LLLDataset(datasets.GeneratorBasedBuilder): """LLL dataset for gene interaction extraction (RE)""" SOURCE_VERSION = datasets.Version(_SOURCE_VERSION) BIGBIO_VERSION = datasets.Version(_BIGBIO_VERSION) BUILDER_CONFIGS = [ BigBioConfig( name="lll_source", version=SOURCE_VERSION, description="LLL source schema", schema="source", subset_id="lll", ), BigBioConfig( name="lll_bigbio_kb", version=BIGBIO_VERSION, description="LLL BigBio schema", schema="bigbio_kb", subset_id="lll", ), ] DEFAULT_CONFIG_NAME = "lll_source" def _info(self) -> datasets.DatasetInfo: if self.config.schema == "source": features = datasets.Features( { "id": datasets.Value("string"), "sentence": datasets.Value("string"), "words": [ { "id": datasets.Value("string"), "text": datasets.Value("string"), "offsets": datasets.Sequence(datasets.Value("int32")), } ], "genic_interactions": [ { "ref_id1": datasets.Value("string"), "ref_id2": datasets.Value("string"), } ], "agents": [ { "ref_id": datasets.Value("string"), } ], "targets": [ { "ref_id": datasets.Value("string"), } ], "lemmas": [ { "ref_id": datasets.Value("string"), "lemma": datasets.Value("string"), } ], "syntactic_relations": [ { "type": datasets.Value("string"), "ref_id1": datasets.Value("string"), "ref_id2": datasets.Value("string"), } ], } ) elif self.config.schema == "bigbio_kb": features = kb_features return datasets.DatasetInfo( description=_DESCRIPTION, features=features, homepage=_HOMEPAGE, license=str(_LICENSE), citation=_CITATION, ) def _split_generators(self, dl_manager) -> List[datasets.SplitGenerator]: urls = _URLS[_DATASETNAME] train_path, train_coref_path, test_path = dl_manager.download_and_extract(urls) return [ datasets.SplitGenerator( name=datasets.Split.TRAIN, gen_kwargs={ "data_paths": [train_path, train_coref_path], "split": "train", }, ), datasets.SplitGenerator( name=datasets.Split.TEST, gen_kwargs={"data_paths": [test_path], "split": "test"}, ), ] def _generate_examples(self, data_paths, split): if self.config.schema == "source": for path in data_paths: with open(path, encoding="utf8") as documents: for document in self._generate_parsed_documents(documents, split): yield document["id"], document elif self.config.schema == "bigbio_kb": uid = it.count(0) for path in data_paths: with open(path, encoding="utf8") as documents: for document in self._generate_parsed_documents(documents, split): document_ = {} document_["id"] = next(uid) document_["document_id"] = document["id"] document_["passages"] = [ { "id": next(uid), "type": BigBioValues.NULL, "text": [document["sentence"]], "offsets": [[0, len(document["sentence"])]], } ] id_to_word = {i["id"]: i for i in document["words"]} document_["entities"] = [] for agent in document["agents"]: word = id_to_word[agent["ref_id"]] document_["entities"].append( { "id": f"{document_['id']}-agent-{word['id']}", "type": "agent", "text": [word["text"]], "offsets": [ [word["offsets"][0], word["offsets"][1]] ], "normalized": [], } ) for agent in document["targets"]: word = id_to_word[agent["ref_id"]] document_["entities"].append( { "id": f"{document_['id']}-target-{word['id']}", "type": "target", "text": [word["text"]], "offsets": [ [word["offsets"][0], word["offsets"][1]] ], "normalized": [], } ) document_["relations"] = [ { "id": next(uid), "type": "genic_interaction", "arg1_id": f"{document_['id']}-agent-{relation['ref_id1']}", "arg2_id": f"{document_['id']}-target-{relation['ref_id2']}", "normalized": [], } for relation in document["genic_interactions"] ] document_["events"] = [] document_["coreferences"] = [] yield document_["document_id"], document_ def _generate_parsed_documents(self, fstream, split): for raw_document in self._generate_raw_documents(fstream): yield self._parse_document(raw_document, split) def _generate_raw_documents(self, fstream): raw_document = [] for line in fstream: if "%" in line: continue elif line.strip(): raw_document.append(line.strip()) elif raw_document: if raw_document: yield raw_document raw_document = [] # needed for last document if raw_document: yield raw_document def _parse_document(self, raw_document, split): document = {} for line in raw_document: key, value = line.split("\t", 1) if key in ["ID", "sentence"]: document[key.lower()] = value elif key in [ "words", "genic_interactions", "agents", "targets", "lemmas", "syntactic_relations", ]: document[key.lower()] = self._parse_elements(value, key) else: raise NotImplementedError() # Needed as testset does not contain agents, targets and genic_interactions (dataset was part of a challenge) if split == "test": document.setdefault("genic_interactions", []) document.setdefault("agents", []) document.setdefault("targets", []) return document def _parse_elements(self, values, type): return [self._parse_element(atom, type) for atom in values.split("\t")] def _parse_element(self, atom, type): # Sorry for that abomination, parses the arguments from atoms like rel(arg1, ..., argn) args = atom.split("(", 1)[1][:-1].split(",") if type == "words": # fix offsets for python slicing return { "id": args[0], "text": args[1].strip("'"), "offsets": [int(args[2]), int(args[3]) + 1], } elif type == "genic_interactions": return {"ref_id1": args[0], "ref_id2": args[1]} elif type == "agents": return {"ref_id": args[0]} elif type == "targets": return {"ref_id": args[0]} elif type == "lemmas": return {"ref_id": args[0], "lemma": args[1].strip("'")} elif type == "syntactic_relations": return {"type": args[0].strip("'"), "ref_id1": args[1], "ref_id2": args[2]}