import sys

import configparser
import time
import re
import csv

import pandas as pd
import numpy as np
from pandas import Series,DataFrame

DIC_NUMS = 5000

class PCFG_gene:
	def __init__(self,structure_df,str_list):
		self.structure_df = structure_df
		self.str_list = str_list

	def PCFG_Pre(self):
		stru_dic = {}
		stru_nums = {}
		for tmp_list in self.str_list:
			if tmp_list == []:
				continue
			tmp_dict = {}
			sums = 0
			for i in range(1,len(tmp_list)):
				str_toparse = tmp_list[i]
				index = re.search(r'-\d+$',str_toparse).span()
				num = int(str_toparse[index[0]+1:])
				sums += num
				tmp_dict[str_toparse[:index[0]]] = num
			stru_dic[tmp_list[0]] = tmp_dict
			stru_nums[tmp_list[0]] = sums
		return stru_dic,stru_nums
		
	def PCFG_gene_list(self,stru_dic,stru_nums):
		passwd_list = {}
		for i in range(len(self.structure_df)):
				passwd_list[self.structure_df.iloc[i][0]] = float(self.structure_df.iloc[i][1])
		result = {}
		cnum = 0
		for key in passwd_list.keys():
			if cnum==234:
				print(1)
			print (cnum)
			cnum += 1
			parse_list = re.findall(r'[A-Z]\d+',key)
			final_freq = passwd_list[key]
			final_list = {'':final_freq}
			for sub_str in parse_list:
				lis = sorted(stru_dic[sub_str].items(),key=lambda item:item[1],reverse=True)
				tmp_dic = {}
				for j in range(int(passwd_list[key]*DIC_NUMS) + 1):  #这里的长度选取有待商榷
					if j < len(lis):
						prob = lis[j][1] * 1.0 / stru_nums[sub_str]
						for s in final_list.keys():
							tmp_dic[lis[j][0] + s] = final_list[s] * prob
				final_list = tmp_dic
			if final_list:
				result[key] = final_list
		result_list = {}
		for k in result.keys():
			for r in result[k].keys():
				result_list[r] = result[k][r]
		res = sorted(result_list.items(),key=lambda item:item[1],reverse=True)
		df = DataFrame(columns=('passwd','prob'))
		for q in range(min(DIC_NUMS,len(res))):
			df.loc[res[q][0]] = [res[q][0],res[q][1]]
		df.to_csv('pcfg_gene4.csv',index = False)
					

if __name__ == '__main__':

	time1 = time.perf_counter()

	#--------------------读文件模块--------------------#
	#读取结构串文件
	structure_df = pd.read_csv('./structure.csv',header=None)
	csv_file = csv.reader(open('./StringFile.csv'))
	str_list = []
	for i in csv_file:
		str_list.append(i)

	time2 = time.perf_counter()
	print (time2 - time1)

	p = PCFG_gene(structure_df,str_list)
	stru_dic,stru_nums = p.PCFG_Pre()
	p.PCFG_gene_list(stru_dic,stru_nums)


	#--------------------分析模块--------------------#

	time3 = time.perf_counter()
	print  (time3 - time2)