def __init__(self, endpoint='http://localhost:5001/qg/api/v1.0/query', cache_path="", use_cache=True): self.endpoint = endpoint self.cache_path = cache_path self.use_cache = use_cache self.cache = {} if self.use_cache: Utils.makedirs(cache_path) self.cache_path = os.path.join(cache_path, "sqg.cache") self.__load_cache()
def __init__(self, endpoint='http://localhost:5000/processQuery', cache_path="", use_cache=True): self.endpoint = endpoint self.cache_path = cache_path self.use_cache = use_cache self.cache = {} if self.use_cache: Utils.makedirs(cache_path) self.cache_path = os.path.join(cache_path, "earl.cache") self.__load_cache()
from common.utility.utils import Utils from common.component.chunker.SENNAChunker import SENNAChunker from common.component.chunker.classifierChunkParser import ClassifierChunkParser from common.component.chunker.goldChunker import GoldChunker import pickle as pk import nltk import os if __name__ == "__main__": test_sentence = "Was winston churchill the prime minister of Selwyn Lloyd?" test_sentence = "What is the hometown of Nader Guirat, where Josef Johansson was born too?" base_dir = "../../" model_dir = os.path.join(base_dir, "models") Utils.makedirs(model_dir) with open('../../data/LC-QUAD/linked_IOB.pk') as data_file: orginal_dataset = pk.load(data_file) dataset = [item[1:] for item in orginal_dataset] idx = int(len(dataset) * 0.9) # idx = -1 train_sents = dataset[:idx] test_sents = dataset[idx + 1:] test_sents_tree = [nltk.chunk.conlltags2tree(item) for item in test_sents] gold_chunker = GoldChunker({item[0]: item[1:] for item in orginal_dataset}) print "\nGold Chunker" print gold_chunker.evaluate(test_sents_tree) SENNA_chunker = SENNAChunker() print "SENNA Chunker" print SENNA_chunker.evaluate(test_sents_tree)