# 分数（15）
class my_CYK(object):
    # TODO: 初始化函数
    def __init__(self, non_ternimal, terminal, rules_prob, start_prob):
        

    # TODO: 返回句子的句法结构，并以树状结构打印出来
    def parse_sentence(self, sentence):



def main(sentence):
    """
    主函数，句法结构分析需要的材料如下
    """
    non_terminal = {'S', 'NP', 'VP', 'PP', 'DT', 'Vi', 'Vt', 'NN', 'IN'}
    start_symbol = 'S'
    terminal = {'sleeps', 'saw', 'boy', 'girl', 'dog', 'telescope', 'the', 'with', 'in'}
    rules_prob = {'S': {('NP', 'VP'): 1.0},
                  'VP': {('Vt', 'NP'): 0.8, ('VP', 'PP'): 0.2},
                  'NP': {('DT', 'NN'): 0.8, ('NP', 'PP'): 0.2},
                  'PP': {('IN', 'NP'): 1.0},
                  'Vi': {('sleeps',): 1.0},
                  'Vt': {('saw',): 1.0},
                  'NN': {('boy',): 0.1, ('girl',): 0.1,('telescope',): 0.3,('dog',): 0.5},
                  'DT': {('the',): 1.0},
                  'IN': {('with',): 0.6, ('in',): 0.4},
                }
    cyk = my_CYK(non_terminal, terminal, rules_prob, start_symbol)
    cyk.parse_sentence(sentence)


# TODO: 对该测试用例进行测试
if __name__ == "__main__":
    sentence = "the boy saw the dog with the telescope"
    main(sentence)

# 分数（3）
# 上面所写的算法的时间复杂度和空间复杂度分别是多少？
# TODO
时间复杂度=O(), 空间复杂度=O()

import jieba
import pandas as pd

# TODO：将entity_list.csv中已知实体的名称导入分词词典
entity_data = pd.read_csv('../data/entity_disambiguation/entity_list.csv', encoding = 'utf-8')


# TODO：对每句句子识别并匹配实体     
valid_data = pd.read_csv('../data/entity_disambiguation/valid_data.csv', encoding = 'gb18030')
    

# TODO：将计算结果存入文件
"""
格式为第一列是需要消歧的语句序号，第二列为多个“实体起始位坐标-实体结束位坐标：实体序号”以“|”分隔的字符串。
样例如下：
[
    [0, '3-6:1008|109-112:1008|187-190:1008'],
    ...
]
"""
pd.DataFrame(result_data).to_csv('../submit/entity_disambiguation_submit.csv', index=False)

# code
# 首先尝试利用开源工具分出实体

import fool  # foolnltk是基于深度学习的开源分词工具，参考https://github.com/rockyzhengwu/FoolNLTK，也可以使用LTP等开源工具
import pandas as pd
from copy import copy


sample_data = pd.read_csv('../data/info_extract/train_data.csv', encoding = 'utf-8', header=0)
y = sample_data.loc[:,['tag']]
train_num = len(sample_data)
test_data = pd.read_csv('../data/info_extract/test_data.csv', encoding = 'utf-8', header=0)
sample_data = pd.concat([sample_data.loc[:, ['id', 'sentence']], test_data])
sample_data['ner'] = None
# TODO: 将提取的实体以合适的方式在‘ner’列中，便于后续使用

import jieba
import jieba.posseg as pseg
import re
import datetime



#提示：可以分析公司全称的组成方式，将“地名”、“公司主体部分”、“公司后缀”区分开，并制定一些启发式的规则
# TODO：建立main_extract，当输入公司名，返回会被统一的简称
def main_extract(company_name,stop_word,d_4_delete,d_city_province):
    """
    company_name  输入的公司名
    stop_word 停用词
    d_city_province 地区
    """

# 简单测试实体统一代码
stop_word,d_city_province = my_initial()
company_name = "河北银行股份有限公司"
# 对公司名提取主体部分，将包含相同主体部分的公司统一为一个实体
company_name = main_extract(company_name,stop_word,d_city_province)
print(company_name)

# TODO：在实体识别中运用统一实体的功能

import fool
import pandas as pd
from copy import copy


sample_data = pd.read_csv('../data/info_extract/train_data.csv', encoding = 'utf-8', header=0)
y = sample_data.loc[:,['tag']]
train_num = len(sample_data)
test_data = pd.read_csv('../data/info_extract/test_data.csv', encoding = 'utf-8', header=0)
sample_data = pd.concat([sample_data.loc[:, ['id', 'sentence']], test_data])
sample_data['ner'] = None
# TODO: 将提取的实体以合适的方式在‘ner’列中并统一编号，便于后续使用

# 提取文本tf-idf特征
# 去除停用词，并转换成tfidf向量。
# 可以使用LTP分词工具，参考：https://ltp.readthedocs.io/zh_CN/latest/
from sklearn.feature_extraction.text import TfidfTransformer  
from sklearn.feature_extraction.text import CountVectorizer  
from pyltp import Segmentor

def get_tfidf_feature():
    segmentor = Segmentor()  # 初始化实例
    segmentor.load_with_lexicon('/ltp_data_v3.4.0/cws.model', '../data/user_dict.txt')  # 加载模型

    
    return tfidf_feature

# -*- coding: utf-8 -*-
from pyltp import Parser
from pyltp import Segmentor
from pyltp import Postagger
import networkx as nx
import pylab
import re

# 投资关系关键词
# 提示：可以结合投资关系的触发词建立有效特征
key_words = ["收购","竞拍","转让","扩张","并购","注资","整合","并入","竞购","竞买","支付","收购价","收购价格","承购","购得","购进",
             "购入","买进","买入","赎买","购销","议购","函购","函售","抛售","售卖","销售","转售"]

postagger = Postagger() # 初始化实例
postagger.load_with_lexicon('/ltp_data_v3.4.0/pos.model', '../data/user_dict.txt')  # 加载模型
segmentor = Segmentor()  # 初始化实例
segmentor.load_with_lexicon('/ltp_data_v3.4.0/cws.model', '../data/user_dict.txt')  # 加载模型
parser = Parser() # 初始化实例
parser.load('/ltp_data_v3.4.0/parser.model')  # 加载模型


def get_parse_feature(s):
    """
    对语句进行句法分析，并返回句法结果
    parse_result：依存句法解析结果
    source：企业实体的词序号
    target：另一个企业实体的词序号
    keyword_pos：关键词词序号列表
    source_dep：企业实体依存句法类型
    target_dep：另一个企业实体依存句法类型
    ...
    （可自己建立新特征，提高分类准确率）
    """


    return features



# LTP中的依存句法类型如下：['SBV', 'VOB', 'IOB', 'FOB', 'DBL', 'ATT', 'ADV', 'CMP', 'COO', 'POB', 'LAD', 'RAD', 'IS', 'HED']

# 汇总词频特征和句法特征
from sklearn.preprocessing import OneHotEncoder

whole_feature = pd.concat([tfidf_feature, parse_feature])
# TODO: 将字符型变量转换为onehot形式

train_x = whole_feature.iloc[:, :train_num]
test_x = whole_feature.iloc[:, train_num:]

# 建立分类器进行分类，使用sklearn中的分类器，不限于LR、SVM、决策树等
from sklearn.ensemble import RandomForestClassifier
from sklearn import preprocessing
from sklearn.model_selection import train_test_split


class RF:
    def __init__(self):
 
    def train(self, train_x, train_y):
        
        return model
        
    def predict(self, model, test_x)

        return predict, predict_prob
    
    
rf = RF()
model = rf.train(train_x, y)
predict, predict_prob = rf.predict(model, test_x)


# TODO: 存储提取的投资关系实体对，本次关系抽取不要求确定投资方和被投资方，仅确定实体对具有投资关系即可
"""
以如下形式存储，转为dataframe后写入csv文件：
[
    [九州通,江中药业股份有限公司],
    ...
]
"""

from py2neo import Node, Relationship, Graph

graph = Graph(
    "http://localhost:7474", 
    username="neo4j", 
    password="person"
)

for v in relation_list:
    a = Node('Company', name=v[0])
    b = Node('Company', name=v[1])
    
    # 本次不区分投资方和被投资方
    r = Relationship(a, 'INVEST', b)
    s = a | b | r
    graph.create(s)
    r = Relationship(b, 'INVEST', a)
    s = a | b | r
    graph.create(s)

# TODO：查询某节点的3层投资关系