BIDI_ML_INFO_EXTRACTION/BiddingKG/dl_dev/labelServer/definition.py

import pickle
import requests
import json

from ipywidgets import widgets
from IPython.display import display,clear_output

import os
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID" 
os.environ["CUDA_VISIBLE_DEVICES"] = ""

def getHbox(entity):
    check = False if entity[5]=="1" else True
    return widgets.HBox([widgets.ToggleButton(
            value=check,
            description='表述错误',
            disabled=False,
            layout=widgets.Layout(width="100px",height="100px"),
                icon='check'
        ),
            widgets.Label(value="表述：",layout=widgets.Layout(width="60px",height="100px")),
            widgets.Textarea(value=getBS(entity),layout=widgets.Layout(width="170px",height="100px")),
            widgets.Label(value="前后文：",layout=widgets.Layout(width="100px",height="100px")),
            widgets.Textarea(value="".join(entity[0]),layout=widgets.Layout(width="170px",height="100px")),
            widgets.Textarea(value="".join(entity[1]),layout=widgets.Layout(width="170px",height="100px")),
            widgets.Textarea(value="".join(entity[2]),layout=widgets.Layout(width="170px",height="100px"))])

def save(object_to_save, path):
    '''
    保存对象
    @Arugs:
        object_to_save: 需要保存的对象

    @Return:
        保存的路径
    '''
    with open(path, 'wb') as f:
        pickle.dump(object_to_save, f)

def load(path):
    '''
    读取对象
    @Arugs:
        path: 读取的路径

    @Return:
        读取的对象
    '''
    with open(path, 'rb') as f:
        object1 = pickle.load(f)
        return object1
    
guardian_base = 'http://127.0.0.1:15010'
myheaders = {'Content-Type': 'application/json'}
source_data_file = "data.pk"


import psycopg2

from DBUtils.PooledDB import PooledDB

pool = None

def getConnection():
    global pool
    if pool is None:
        pool = PooledDB(psycopg2, 5,5,dbname="article_label", host="192.168.2.101",user="postgres",password="postgres",port="5432")
    return pool.connection()

def make(index_,source_data):
    user = {
                "id": source_data[index_][0],
                "content":source_data[index_][1]
        }
    _resp = requests.post(guardian_base + '/article_extract', json=user, headers=myheaders, verify=True)
    return json.loads(_resp.content.decode("utf-8"))["success"] is True

BS_dic = {"org":{"0":"角色-招标人","1":"角色-代理人","2":"角色-中标/第一候选人","3":"角色-第二候选人","4":"角色-第三候选人","5":"角色-无"},
         "company":{"0":"角色-招标人","1":"角色-代理人","2":"角色-中标/第一候选人","3":"角色-第二候选人","4":"角色-第三候选人","5":"角色-无"},
         "money":{"0":"金额-招标金额","1":"金额-中投标金额","2":"金额-其他金额"},
         "person":{"0":"联系人-非目标联系人","1":"联系人-招标联系人","2":"联系人-代理联系人","3":"联系人-联系人"}}

def getBS(entity):
    return BS_dic[entity[3]][entity[4]]

def getEntitys(index_,source_data):
    id = source_data[index_][0]
    conn = getConnection()
    cursor = conn.cursor()
    sql = " select B.tokens,A.entity_text,A.entity_type,A.label,A.handlabel,A.entity_id,A.begin_index,A.end_index,A.values from entity_mention A,sentences B where A.doc_id=B.doc_id and A.sentence_index=B.sentence_index and A.label !='None' "+\
    " and B.doc_id='"+id+"' order by A.label,A.entity_type "
    cursor.execute(sql)
    rows = cursor.fetchall()
    data = []
    for row in rows:
        tokens = row[0]
        entity_text = row[1]
        entity_type = row[2]
        label = row[3]
        handlabel = row[4]
        entity_id = row[5]
        begin_index = row[6]
        end_index = row[7]
        values = row[8]
        prob = values[1:-1].split(",")[int(label)]
        if float(prob)<0.5:
            continue
        span = spanWindow(tokens,begin_index,end_index,10)
        data.append([span[0],span[1],span[2],entity_type,label,handlabel,entity_id])
    conn.close()
    return data

def spanWindow(tokens,begin_index,end_index,size):
    '''
    @summary:取得某个实体的上下文词汇
    @param:
        tokens:句子分词list
        begin_index:实体的开始index
        end_index:实体的结束index
        size:左右两边各取多少个词
    @return: list,实体的上下文词汇
    '''  
    length_tokens = len(tokens)
    if begin_index>size:
        begin = begin_index-size
    else:
        begin = 0
    if end_index+size<length_tokens:
        end = end_index+size+1
    else:
        end = length_tokens
    result = []
    result.append(tokens[begin:begin_index])
    result.append(tokens[begin_index:end_index+1])
    result.append(tokens[end_index+1:end])
    return result

def getCodeName(index_,source_data):
    id = source_data[index_][0]
    conn = getConnection()
    cursor = conn.cursor()
    sql = " select code,name from articles_processed where id='"+id+"' "
    cursor.execute(sql)
    rows = cursor.fetchall()
    conn.close()
    if len(rows)>0:
        return rows[0][0],rows[0][1]
    else:
        return "",""

def saveData(datas,out_code,begin_index,source_data,out_name,out_vbox):
    if out_code.value=="" and out_name.value=="":
        print("请标注编号名称")
        return 1
    conn = getConnection()
    cursor = conn.cursor()
    sql = " update articles_processed set code='"+out_code.value+"',name='"+out_name.value+"' where id='"+source_data[begin_index][0]+"'"
    cursor.execute(sql)
    
    for i in range(len(datas)):
        handlabel = "0" if out_vbox.children[i].children[0].value else "1"
        if handlabel == "0":
            sql = " update entity_mention set handlabel='"+handlabel+"' where entity_id='"+datas[i][6]+"' and entity_type='"+datas[i][3]+"'"
            cursor.execute(sql)
    conn.commit()
    conn.close()
    return 0