def __init__(self, time_gap):
self._graph = Graph(
f"http://{ApplicationConfig.get_neo4j_ip()}:{ApplicationConfig.get_neo4j_port()}/browser",
user=ApplicationConfig.get_neo4j_user(),
password=ApplicationConfig.get_neo4j_password())
self._node_matcher = NodeMatcher(self._graph)
self._rel_matcher = RelationshipMatcher(self._graph)
query = 'MATCH (n) WHERE EXISTS(n.date) RETURN DISTINCT n.date AS date'
date = self._graph.run(query).to_data_frame()
date = date.sort_values(by='date')
date['date'] = date.apply(
lambda x: datetime.strptime(x['date'], '%Y-%m-%d').date(), axis=1)
self._date_integrate(date, time_gap)
def __init__(self, sentence):
config = ApplicationConfig.get_corenlp_config()
self._parser = CoreNLPParser(url=f"http://{config['host']}:{config['port']}")
self._dependency = CoreNLPDependencyParser(url=f"http://{config['host']}:{config['port']}")
sentence = sentence.replace(' ', ' ')
sentence = sentence.replace('.', '')
self._load(sentence)
self.original = sentence
def __init__(self):
self._key = ['subject', 'predicate', 'object']
self._NodeHeader = [
'label', 'date', 'name', 'newsId', 'number', 'source'
]
self._RelHeader = [
'start', 'type', 'end', 'relation', 'date', 'newsId', 'number',
'source'
]
self._path = ApplicationConfig.get_neo4j_svo_output_path()
import pandas as pd
import numpy as np
import pickle
import string
from os import listdir
from datetime import datetime
from tqdm import tqdm
import SVO_final as SVO
import re
from cathay.config import ApplicationConfig
import boto3
from multiprocessing import Pool
import nltk
import sys
aws_nlu_config = ApplicationConfig.get_aws_nlu_config()
comprehend = boto3.client(aws_access_key_id=aws_nlu_config['access_key'],
aws_secret_access_key=aws_nlu_config['secret_key'],
service_name='comprehend',
region_name=aws_nlu_config['region'])
def title_preprocess(sent, comprehend):
if sent.find('-') == 8 and sent[:6] == 'UPDATE':
sent = sent[9:]
if sent.find('-') == 8 and sent[:6] == 'WRAPUP':
sent = sent[9:]
if sent.find('-') == 3 and sent[:3] == 'RPT':
sent = sent[5:]
while sent.find('-') != -1 and sent[:sent.find('-')].isupper():
sent = sent[sent.find('-') + 1:]
def __init__(self, folder):
self.path = ApplicationConfig.get_neo4j_svo_output_path() + folder
self._save()
self._merge()
def __init__(self):
self._graph = Graph(
f"http://{ApplicationConfig.get_neo4j_ip()}:{ApplicationConfig.get_neo4j_port()}/browser",
user=ApplicationConfig.get_neo4j_user(),
password=ApplicationConfig.get_neo4j_password())