[ Python 常見問題 ] Python HTMLParser: UnicodeDecodeError

Source From Here 
Question 
I'm using HTMLParser to parse pages I pull down with urllib, and am coming across UnicodeDecodeError exceptions when passing some to HTMLParser. The information is there if I just print it out. 

view plaincopy to clipboardprint?

1. from HTMLParser import HTMLParser  
2. import urllib  
3. import chardet  
4.   
5. class search_youtube(HTMLParser):  
6.   
7.     def __init__(self, search_terms):  
8.         HTMLParser.__init__(self)  
9.         self.track_ids = []  
10.         for search in search_terms:  
11.             self.__in_result = False  
12.             search = urllib.quote_plus(search)  
13.             query = 'http://youtube.com/results?search_query='  
14.             page = urllib.urlopen(query + search).read()  
15.             try:  
16.                 self.feed(page)  
17.             except UnicodeDecodeError:  
18.                 encoding = chardet.detect(page)['encoding']  
19.                 if encoding != 'unicode':  
20.                     page = page.decode(encoding)  
21.                     page = page.encode('ascii', 'ignore')  
22.                 self.feed(page)  
23.                 print 'success'  
24.   
25. searches = ['telepopmusik breathe']  
26. results = search_youtube(searches)  
27. print results.track_ids  

Here's the output: 

view plaincopy to clipboardprint?

1. Traceback (most recent call last):  
2.   File "test.py", line 27, in   
3.     results = search_youtube(searches)  
4.   File "test.py", line 23, in __init__  
5.     self.feed(page)  
6.   File "/usr/lib/python2.6/HTMLParser.py", line 108, in feed  
7.     self.goahead(0)  
8.   File "/usr/lib/python2.6/HTMLParser.py", line 148, in goahead  
9.     k = self.parse_starttag(i)  
10.   File "/usr/lib/python2.6/HTMLParser.py", line 252, in parse_starttag  
11.     attrvalue = self.unescape(attrvalue)  
12.   File "/usr/lib/python2.6/HTMLParser.py", line 390, in unescape  
13.     return re.sub(r"&(#?[xX]?(?:[0-9a-fA-F]+|\w{1,8}));", replaceEntities, s)  
14.   File "/usr/lib/python2.6/re.py", line 151, in sub  
15.     return _compile(pattern, 0).sub(repl, string, count)  
16. UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 1: ordinal not in range(128)  

How-To 
It is UTF-8, indeed. This works: 

view plaincopy to clipboardprint?

1. from HTMLParser import HTMLParser  
2. import urllib  
3.   
4. class search_youtube(HTMLParser):  
5.   
6.     def __init__(self, search_terms):  
7.         HTMLParser.__init__(self)  
8.         self.track_ids = []  
9.         for search in search_terms:  
10.             self.__in_result = False  
11.             search = urllib.quote_plus(search)  
12.             query = 'http://youtube.com/results?search_query='  
13.             connection = urllib.urlopen(query + search)  
14.             encoding = connection.headers.getparam('charset')  
15.             if encoding:  
16.                 page = connection.read().decode(encoding)  
17.             else:  
18.                 page = connection.read()  
19.             self.feed(page)  
20.             print 'success'  
21.   
22. searches = ['telepopmusik breathe']  
23. results = search_youtube(searches)  
24. print results.track_ids  

You don't need chardet, Youtube are not morons, they actually send the correct encoding in the header.