Kamitani dowloading fails

[GaelVaroquaux]

$ python kamitani.py
/home/varoquau/dev/scikit-learn/scikits/__init__.py:1: UserWarning: Module dateutil was already imported from /home/varoquau/dev/matplotlib/lib/dateutil/__init__.pyc, but /usr/lib/python2.7/dist-packages is being added to sys.path
  __import__('pkg_resources').declare_namespace(__name__)
Downloading data from public_beta_20100515.zip ...
Traceback (most recent call last):
  File "kamitani.py", line 54, in 
    dataset = datasets.fetch_kamitani()
  File "/home/varoquau/dev/nisl/tutorial/nisl/datasets.py", line 502, in fetch_kamitani
    force_download=force_download)
  File "/home/varoquau/dev/nisl/tutorial/nisl/datasets.py", line 219, in fetch_dataset
    data = urllib2.urlopen(req)
  File "/usr/lib/python2.7/urllib2.py", line 126, in urlopen
    return _opener.open(url, data, timeout)
  File "/usr/lib/python2.7/urllib2.py", line 392, in open
    protocol = req.get_type()
  File "/usr/lib/python2.7/urllib2.py", line 254, in get_type
    raise ValueError, "unknown url type: %s" % self.__original
ValueError: unknown url type: public_beta_20100515.zip

[AlexandreAbraham]

I remember the story of this one. There is in fact no URL for Kamitani dataset because it requires a registration. I will temporarily put an error message but I see 4 options :

• use a pre-created account to download it
• ask the user for registration information and use them to get the dataset
• create a mirror of the dataset
• tell the user what to do (registration, download and move to the right directory)

[GaelVaroquaux]

On Fri, Jun 22, 2012 at 01:50:48PM -0700, Alexandre Abraham wrote:

• use a pre-created account to download it

That option.

G

[AlexandreAbraham]

My bad, the download system seems to have changed... The dataset is now available at : http://brainliner.jp/data/brainliner-admin/Reconstruct Good thing is, no authentication is required. Plus, file format seems to be the same (according to public_beta.pdf).I will try to make a downloader for it.

[AlexandreAbraham]

Okay, I'm desperately trying to download through the new system but I did not manage to do it. Plus, the data format has changed so it will require more work. I will try to finish all other datasets before fixing this one.