Difficulties uploading datasets #292

Activity

1. joaquinvanschoren commented on Aug 13, 2017

   

   joaquinvanschoren

   on Aug 13, 2017

   ContributorAuthor

   More actions

   Weirdly, this also fails. The test server fails to parse the ARFF file?

   dataset = oml.datasets.OpenMLDataset(
               name="UploadTestWithURL", version=1, description="test",
               format="ARFF",
               url="https://www.openml.org/data/download/61/dataset_61_iris.arff")
   dataset.publish()
   

   ---------------------------------------------------------------------------
   OpenMLServerException                     Traceback (most recent call last)
   <ipython-input-58-99a1903d7914> in <module>()
         9             format="ARFF",
        10             url="https://www.openml.org/data/download/61/dataset_61_iris.arff")
   ---> 11 dataset.publish()
        12 
        13 
   
   /Users/joa/anaconda/lib/python3.5/site-packages/openml/datasets/dataset.py in publish(self)
       392 
       393         return_value = _perform_api_call("/data/", file_dictionary=file_dictionary,
   --> 394                                          file_elements=file_elements)
       395 
       396         self.dataset_id = int(xmltodict.parse(return_value)['oml:upload_data_set']['oml:id'])
   
   /Users/joa/anaconda/lib/python3.5/site-packages/openml/_api_calls.py in _perform_api_call(call, data, file_dictionary, file_elements, add_authentication)
        49     if file_dictionary is not None or file_elements is not None:
        50         return _read_url_files(url, data=data, file_dictionary=file_dictionary,
   ---> 51                                file_elements=file_elements)
        52     return _read_url(url, data)
        53 
   
   /Users/joa/anaconda/lib/python3.5/site-packages/openml/_api_calls.py in _read_url_files(url, data, file_dictionary, file_elements)
        95     response = requests.post(url, data=data, files=file_elements)
        96     if response.status_code != 200:
   ---> 97         raise _parse_server_exception(response)
        98     if 'Content-Encoding' not in response.headers or \
        99             response.headers['Content-Encoding'] != 'gzip':
   
   OpenMLServerException: Error parsing dataset ARFF file
   
2. joaquinvanschoren commented on Aug 14, 2017

   

   joaquinvanschoren

   on Aug 14, 2017

   Last edited by joaquinvanschoren

   ContributorAuthor

   More actions

   Sorry, 3rd question on this topic: how can I upload a dataset that is locally stored?
   I tried stating the file path with 'data_file', and while an OpenMLDataset object is created, during upload in fails with the same parsing error as above.
3. joaquinvanschoren

   changed the title [-]Outdated URL makes tests fail[/-] [+]Difficulties uploading datasets[/+] on Aug 14, 2017
4. janvanrijn commented on Aug 14, 2017

   

   janvanrijn

   on Aug 14, 2017

   Member

   More actions

   Answers to the questions:
   a) This is done on purpose. We wanted an ARFF from the web, and this one seemed appropriate. Connection between openml and expdb is coincidental.
   b) I would guess that the arff of iris (and many other UCI arffs) do not comply with the arff standards required by the openml server (it mixes tabs and spaces)
   c) cannot answer this ATM
5. joaquinvanschoren commented on Aug 14, 2017

   

   joaquinvanschoren

   on Aug 14, 2017

   Last edited by joaquinvanschoren

   ContributorAuthor

   More actions

   a) I meant that that kuleuven server is no longer available. The URL gives a 404.
   b) Fair enough. It is really hard to see this though :)
   c) Works now.
6. joaquinvanschoren commented on Aug 14, 2017

   

   joaquinvanschoren

   on Aug 14, 2017

   ContributorAuthor

   More actions

   All good, except for that url in the unit test.
7. mfeurer commented on Aug 15, 2017

   

   mfeurer

   on Aug 15, 2017

   Collaborator

   More actions

   Hm, why don't the unit tests fail then? https://travis-ci.org/openml/openml-python/builds/264404343
8. mfeurer

   added

   bug

   on Sep 29, 2017
9. joaquinvanschoren commented on Apr 17, 2018

   

   joaquinvanschoren

   on Apr 17, 2018

   ContributorAuthor

   More actions

   The error message has changed, but this still looks a bug.

   dataset = openml.datasets.OpenMLDataset(
   name="iristest", version=1, description="test",
   format="ARFF",
   url="http://expdb.cs.kuleuven.be/expdb/data/uci/nominal/iris.arff")
   dataset.publish()
   

   ---------------------------------------------------------------------------
   ExpatError                                Traceback (most recent call last)
   ~/anaconda3/lib/python3.6/site-packages/openml/_api_calls.py in _parse_server_exception(response, url)
       130     try:
   --> 131         server_exception = xmltodict.parse(response.text)
       132     except:
   
   ~/anaconda3/lib/python3.6/site-packages/xmltodict.py in parse(xml_input, encoding, expat, process_namespaces, namespace_separator, disable_entities, **kwargs)
       329     else:
   --> 330         parser.Parse(xml_input, True)
       331     return handler.item
   
   ExpatError: no element found: line 1, column 0
   
   During handling of the above exception, another exception occurred:
   
   OpenMLServerError                         Traceback (most recent call last)
   <ipython-input-20-fa1103f9ad92> in <module>()
   ----> 1 dataset.publish()
   
   ~/anaconda3/lib/python3.6/site-packages/openml/datasets/dataset.py in publish(self)
       423 
       424         return_value = _perform_api_call("/data/", file_dictionary=file_dictionary,
   --> 425                                          file_elements=file_elements)
       426 
       427         self.dataset_id = int(xmltodict.parse(return_value)['oml:upload_data_set']['oml:id'])
   
   ~/anaconda3/lib/python3.6/site-packages/openml/_api_calls.py in _perform_api_call(call, data, file_dictionary, file_elements, add_authentication)
        50     if file_dictionary is not None or file_elements is not None:
        51         return _read_url_files(url, data=data, file_dictionary=file_dictionary,
   ---> 52                                file_elements=file_elements)
        53     return _read_url(url, data)
        54 
   
   ~/anaconda3/lib/python3.6/site-packages/openml/_api_calls.py in _read_url_files(url, data, file_dictionary, file_elements)
        96     response = requests.post(url, data=data, files=file_elements)
        97     if response.status_code != 200:
   ---> 98         raise _parse_server_exception(response, url=url)
        99     if 'Content-Encoding' not in response.headers or \
       100             response.headers['Content-Encoding'] != 'gzip':
   
   ~/anaconda3/lib/python3.6/site-packages/openml/_api_calls.py in _parse_server_exception(response, url)
       133         raise OpenMLServerError(('Unexpected server error. Please '
       134                                  'contact the developers!\nStatus code: '
   --> 135                                  '%d\n' % response.status_code) + response.text)
       136 
       137     code = int(server_exception['oml:error']['oml:code'])
   
   OpenMLServerError: Unexpected server error. Please contact the developers!
   Status code: 500
   
10. mfeurer commented on Jun 8, 2018

    

    mfeurer

    on Jun 8, 2018

    Collaborator

    More actions

    @joaquinvanschoren did you try the latest dataset upload? Is it the way you expect it? If yes, can we close this?
11. joaquinvanschoren commented on Jun 8, 2018

    

    joaquinvanschoren

    on Jun 8, 2018 via email

    ContributorAuthor

    More actions

    I'll try to test it over the weekend

    On Fri, 8 Jun 2018 at 13:41 Matthias Feurer ***@***.***> wrote: @joaquinvanschoren <https://github.com/joaquinvanschoren> did you try the latest dataset upload? Is it the way you expect it? If yes, can we close this? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#292 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABpQV7TdPSmN4rpWUsAzfWCELw1E4Gv-ks5t6mLbgaJpZM4O11E1> .

    -- Thank you, Joaquin
12. mfeurer commented on Nov 16, 2018

    

    mfeurer

    on Nov 16, 2018

    Collaborator

    More actions

    @joaquinvanschoren did you recently try uploading a dataset? Can this be closed?
13. mfeurer commented on Feb 18, 2019

    

    mfeurer

    on Feb 18, 2019

    Collaborator

    More actions

    Closing this due to inactivity. If this is still an issue with the recent 0.8 release, please reopen.
14. mfeurer

    closed this as completedon Feb 18, 2019