###Third code of Transn Internship

def detect_contains_term(input_src, input_target):

detect_result = {}

# print('原文不存在')

data_contains_src = dot_term_info[dot_term_info['src_text'].str.contains(input_src)] # 查找src_text中包含input_src字符串的行

# print(len(data_contains_src))

if len(data_contains_src) > 0: # contains包含字符串查找

# 包含

detect_result['term_input_result'] = '存在相似术语'

data_contains_src = data_contains_src.drop_duplicates() # 去重

# data_contains_src_tran = pd.unique(data_contains_src['target_text'])

data_contains_src_tran = data_contains_src['target_text'].to_numpy() # 转为numpy格式，便于后续求相似度

# 求max_distance

distance = []

for contains_target_text in data_contains_src_tran:

distance.append(cal_distance(contains_target_text, input_target))

max_distance = max(distance)

max_distance_target = data_contains_src_tran[distance.index(max_distance)] # max_distance对应的译文

max_distance_target_src = data_contains_src[data_contains_src['target_text'] == max_distance_target]['src_text'].tolist() # max_distance对应的译文对应的原文，可能有重复

max_distance_target_term_id = data_contains_src[data_contains_src['target_text'] == max_distance_target]['term_id'].tolist() # max_distance对应的译文对应的term_id，可能有重复

# print(max_distance_target_src, max_distance_target)

detect_result['term_id'] = [i for i in max_distance_target_term_id]

detect_result['term_tran_text'] = max_distance_target

# input_src与取出的src_text的distance取average

src_distance = 0

for j in max_distance_target_src:

src_distance += cal_distance(input_src, j)

src_distance = src_distance/len(max_distance_target_src)

# 计算正确率

acc_pro = (max_distance+src_distance)/2

detect_result['term_ratio'] = acc_pro

else:

# 不包含

detect_result['term_input_result'] = '不存在相似的术语'

detect_result['term_id'] = ''

detect_result['term_tran_text'] = ''

detect_result['term_ratio'] = ''

# acc_pro = 'Non_existent,Non_contains,to be continue...'

return detect_result

###Third code of Transn Internship

def detect_match_term(input_target, data_match_src):

detect_result = {}

# print('原文存在')

data_match_src_tran = pd.unique(data_match_src['target_text']) # 取出原文对应的译文的行，去重

# [2.1.1] 计算input_target和target_text之间的max_distance

distance = []

for target_text in data_match_src_tran:

distance.append(cal_distance(target_text, input_target))

max_distance = max(distance) # when max_distance==1,input_target exist,else not exist

if max_distance == 1:

# print('原文存在')

detect_result['term_input_result'] = '存在一样术语'

else:

# print('译文不存在')

detect_result['term_input_result'] = '存在译文不一样术语'

max_distance_target = data_match_src_tran[distance.index(max_distance)] # 取出max_distance对应的译文

max_distance_target_term_id = dot_term_info.loc[dot_term_info['target_text'] == max_distance_target]['term_id'].tolist() #取出max_distance对应的译文的term_id

detect_result['term_id'] = [i for i in max_distance_target_term_id]

detect_result['term_tran_text'] = max_distance_target

# print('术语库中存在的原文及与输入译文最相近对应的译文为：', input_src, max_distance_target)

# max_distance_target

# [2.1.2]

conditional_pro = len(data_match_src.loc[data_match_src['target_text'] == max_distance_target]) / len( data_match_src) # 计算条件概率

# [2.1.3]

acc_pro = (max_distance + conditional_pro) / 2 # 计算正确率

detect_result['term_ratio'] = acc_pro

return detect_result

###Third code of Transn Internship

import pandas as pd
import jieba
from utils.distance_util import cal_distance

# term_info

dot_term_info = pd.read_csv(r'C:\Users\Administrator\Desktop\艾佳\Internship\Detect_term\dot_term_info.csv',

names=['term_id','src_text', 'target_text'], na_filter=False)

# dot_term_info.info()

def detect_term(input_src, input_target):

# [2.1]查找原文是否存在

data_match_src = dot_term_info.loc[lambda x: x['src_text'] == input_src] # 精确查找
# print(len(data_match_src))

if len(data_match_src) >= 1:

# print('原文存在')

detect_result = detect_match_term(input_target, data_match_src)

else:

# print('原文不存在，将查找包含原文字符串的术语')

detect_result = detect_contains_term(input_src, input_target)

return detect_result