I got it working with a utf-8 query by adding an Accept-Charset header to the
request. I used the 'Tamper Data' add-on to Firefox to view all the request
headers being sent by the browser. I added all the same headers to the Python
request and it worked. Then I took out the headers until I found the needed
one. Here is a stripped-down version of your code that posts a word encoded in
utf-8 and gets the correct response. I also changed the post parameters a
little to match what I am seeing in my browser:
import re, urllib, urllib2
__where = [ re.compile(r'name=\"q\">([^<]*)'),
re.compile(r'td bgcolor=white>([^<]*)'),
re.compile(r'td bgcolor=white class=s><div
style=padding:10px;>([^<]*)'),
re.compile(r'<\/strong><br>([^<]*)')
]
phrase = 'ent\xc3\xa3o'
params = urllib.urlencode( { 'doit' : 'done',
'tt' : 'urltext',
'trtext' : phrase,
'intl' : 1,
'lp' : 'pt_en' } )
print "URL encoding ", params
req = urllib2.Request('http://world.altavista.com/babelfish/tr')
req.add_header('Accept-Charset', 'ISO-8859-1,utf-8;q=0.7,*;q=0.7')
response = urllib2.urlopen(req, params)
html = response.read()
for regex in __where:
match = regex.search(html)
if match:
print match.group(1)
break
else:
print "ERROR MATCHING"
print html
Kent
Kent Johnson wrote:
> OK this is actually starting to make sense :-) Here is what I think is
> happening:
>
> You get different results in the IDE and the console because they are using
> different encodings. The IDE is using utf-8 so the params are encoded in
> utf-8. The console is using latin-1 and you get encoded latin-1 params.
>
> When you use babelfish from the browser it gets a page in utf-8 and sends the
> parameters back the same way, but probably with a header saying it is utf-8.
> When you use urllib you don't tell it the encoding so it is assuming latin-1,
> that's why the interpreter version works.
>
> So in your GUI version if you get utf-8 from the GUI, you can convert it to
> latin-1 by
> phrase.decode('utf-8').encode('latin-1') as long as your text can be
> expressed in latin-1. If you need utf-8 then you have to figure out how to
> tell babelfish that you are sending utf-8.
>
> Kent
>
> PS please reply to the list not to me personally.
>
> Jorge Louis de Castro wrote:
>
>>Thanks again,
>>
>>I'm sorry to be such a PITB but this is driving me insane! the code
>>below easily connects to babelfish and returns a translated string.
>>
>>__where = [ re.compile(r'name=\"q\">([^<]*)'),
>> re.compile(r'td bgcolor=white>([^<]*)'),
>> re.compile(r'td bgcolor=white class=s><div
>>style=padding:10px;>([^<]*)'),
>> re.compile(r'<\/strong><br>([^<]*)')
>>
>>def clean(text):
>> return ' '.join(string.replace(text.strip(), "\n", ' ').split())
>>
>>def translateByCode(phrase, from_code, to_code):
>> phrase = clean(phrase)
>> params = urllib.urlencode( { 'BabelFishFrontPage' : 'yes',
>> 'doit' : 'done',
>> 'urltext' : phrase,
>> 'lp' : from_code + '_' + to_code } )
>> print "URL encoding ", params
>> try:
>> response =
>>urllib.urlopen('http://world.altavista.com/babelfish/tr', params)
>> except IOError, what:
>> print "ERRROR TRANSLATING ", what
>> except:
>> print "Unexpected error:", sys.exc_info()[0]
>>
>> html = response.read()
>> for regex in __where:
>> match = regex.search(html)
>> if match: break
>> if not match: print "ERROR MATCHING"
>> return clean(match.group(1))
>>
>>if __name__ == '__main__':
>> print translateByCode('então', 'pt', 'en')
>>
>>If I run this through the Run option on the IDE I get the following output:
>>
>>URL encoding doit=done&urltext=ent%C3%A3o&BabelFishFrontPage=yes&lp=pt_en
>>então
>>então
>>
>>If I import this module on the interpreter and then call
>>
>>print translateByCode('então', 'en', 'pt')
>>
>>I get:
>>
>>URL encoding doit=done&urltext=ent%E3o&BabelFishFrontPage=yes&lp=pt_en
>>then
>>then
>>
>>Now the urllib encoding of the urltext IS different ("ent%C3%A3o" VS
>>"ent%E3o") even though I'm passing the same stuff!
>>And this works fine except when I use special characters and I don't
>>know how to use the utf-8 encoding to get this working -i know altavista
>>uses utf-8 because they also translate chinese.
>>
>>Thanks again and sorry for the blurb but i ran out of solutions for this
>>one.
>>
>>
>>
>
>
> _______________________________________________
> Tutor maillist - Tutor@python.org
> http://mail.python.org/mailman/listinfo/tutor
>
_______________________________________________
Tutor maillist - Tutor@python.org
http://mail.python.org/mailman/listinfo/tutor
