urllib ??

[Rainer]

with the help of alan, we finally have the solution for this problem:
'Submit', 'Process query'),
has to be set as a name/value pair....
this is it

rainer





In article <3DC688E4.BD46E2FB at hotmail.com>, alanmk at hotmail.com says...
>
>Rainer wrote:
>> 
>> hi alan
>> this is the problem, maybe i put it the wrong way.
>> i DO NOT get the results from the server. the thing that is stored in
>> 'results', is the 'query-webpage' with the values filled in, not the
>> results page.
>
>OK, I think I see your peoblem now.
>
>You're submitting data to the server almost correctly, but the server is
>just returning the input form to you, as if you had not filled out all
>of the fields correctly.
>
>And I think that this is one of the things you are doing wrong: not
>filling in the fields correctly. The problem lines are these
>
>    d['.cgifields']='ContAllHits'
>    d['.cgifields']='SortByIde'
>    d['.cgifields']='StructAllHits'
>    d['.cgifields']='Map2Mismatch'
>    d['.cgifields']='Dbase3d'
>
>HTTP allows for multiple values for the same field in a POST request.
>However, python dictionaries do NOT: python dicts can have only one
>value for every key. Try the following at an interactive prompt
>
>>>>d = {}
>>>>d['.cgifields']='ContAllHits'
>>>>d
>{'.cgifields': 'ContAllHits'}
>>>>d['.cgifields']='SortByIde'
>>>>d
>{'.cgifields': 'SortByIde'}
>
>You're expecting it to maintain multiple values for the key
>".cgifields", but it doesn't: the value gets overwritten each time,
>meaning that you are submitting POST data that only includes the value
>'.cgifields'='Dbase3d' (the last assignment in the list), and failing to
>supply values for the (presumably necessary) 'ContAllHits', 'SortByIde',
>'StructAllHits', 'Map2Mismatch'.
>
>The solution is to pass a list of tuples to "urlencode" instead of a
>dictionary, like so
>
>#=-----------------------------------------------------
>
>def query2_polyphen():
>
>    d=[
>    ('InputQuerySequence', '>GGG \012 sequence'),
>    ('Var1', 'C'),
>    ('InputQueryPosition', '282'),
>    ('Var2', 'Y'),
>    ('Dbase3d', 'pqs'),
>    ('SortByIde', '1'),
>    ('Map2Mismatch', '0'),
>    ('StructAllHits', '0'),
>    ('ContAllHits', '0'),
>    ('MinHitLen', '100'),
>    ('MinHitIde', '0.5'),
>    ('MaxHitGaps', '20'),
>    ('ContThresh', '6'),
>    ('.cgifields', 'ContAllHits'),
>    ('.cgifields', 'SortByIde'),
>    ('.cgifields', 'StructAllHits'),
>    ('.cgifields', 'Map2Mismatch'),
>    ('.cgifields', 'Dbase3d'),
>    ]
>    query=urllib.urlencode(d)
>    url = 'http://tux.embl-heidelberg.de/ramensky/polyphen.cgi'
>    results = urllib.urlopen(url,query).read()
>    open ('rainer.html','w').write(results)
>#=-----------------------------------------------------
>
>However, I also notice, by manually submitting your form data using a
>browser, that the data you are submitting give rise to an error. I don't
>know enough about molecular biology to know what the error is, but I
>suggest that you get your submitted data right manually before trying to
>submit it automatically.
>
>Also, you might send an email to the provider of the polyphen service
>asking them to somehow make it possible for automated submission scripts
>to recognise when a query has failed, rather than simply sending the
>form again. A return HTTP status value in the 4xx series might be
>suitable, e.g. of the "404 - file not found" variety.
>
>HTH,
>
>alan kennedy
>-----------------------------------------------------
>check http headers here: http://xhaus.com/headers
>email alan:              http://xhaus.com/mailto/alan