linshan
/
EnglishPal
forked from mrlan/EnglishPal
1
0
Fork 0
Code Pull Requests Activity

1

Bug493-GongKeCheng
田其鹭 2023-05-14 21:17:24 +08:00
parent a5c792b782
commit 654fd52c44
1 changed files with 1 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
app/wordfreqCMD.py
View File

@ -37,14 +37,6 @@ def file2str(fname):#文件转字符
f.close() #关闭 f.close() #关闭
return s return s
def str2chararray(str): # 字符串转字符数组
chararray = []
for i in range(len(str)):
chararray.append(str[i])
return chararray
def remove_punctuation(s): # 这里是s是形参 (parameter)。函数被调用时才给s赋值。 def remove_punctuation(s): # 这里是s是形参 (parameter)。函数被调用时才给s赋值。
# carr = str2chararray(s) # 字符串转字符数组 # carr = str2chararray(s) # 字符串转字符数组
@ -52,22 +44,9 @@ def remove_punctuation(s): # 这里是s是形参 (parameter)。函数被调用
special_characters = '&_~=+[]%^@.,?!:;#()"“”—‘’|/\\<>{}' # 把里面的字符都去掉 special_characters = '&_~=+[]%^@.,?!:;#()"“”—‘’|/\\<>{}' # 把里面的字符都去掉
special_words = ('$lt', '$gt', '$') # 特殊词汇 special_words = ('$lt', '$gt', '$') # 特殊词汇
special_characters = '\_©~<=>+-/[]*&$%^@.,?!:;#()"“”—‘’{}|' # 把里面的字符都去掉
for c in special_characters: for c in special_characters:
s = s.replace(c, ' ') # 防止出现把 apple,apple 移掉逗号后变成 appleapple 情况 s = s.replace(c, ' ') # 防止出现把 apple,apple 移掉逗号后变成 appleapple 情况
carr = str2chararray(s) # 字符串转字符数组
for i, value in enumerate(carr):
if value == '&': # 遍历替换
carr[i] = '\''
print('sss' + s)
for j in range(1,5):
if carr[i+j] in [' ','3','9']:
carr[i + j]=''
s = ''.join(carr) # 字符数组转字符串
print('sss'+s)
for w in special_words: # 替换字符串中的剩余特殊字符
s = s.replace(w, ' ')
s = s.replace('--', ' ') s = s.replace('--', ' ')
s = s.strip() # 去除前后的空格 s = s.strip() # 去除前后的空格