use python to download and unzip files instead of wget, unzip command

[CharlesMei]

this fixes #2237 #1559, for windows user, it is not convenient to use wget, mkdir, unzip like linux user.
we can use pure python code to download and unzip dataset, which works for both windows, linux/unix.

[piiswrong]

is zipfile available by default?

[CharlesMei]

Yes, zipfile is in python standard library.

[hjk41]

confirmed. It works on Python 2.7.10 and 3.5.1

On Sat, Jun 4, 2016 at 2:53 AM, CharlesMei notifications@github.com wrote:

In example/image-classification/train_cifar10.py
#2325 (comment):

 os.chdir(data_dir)
 if (not os.path.exists('train.rec')) or \
    (not os.path.exists('test.rec')) :

•    os.system("wget http://webdocs.cs.ualberta.ca/~bx3/data/cifar10.zip")
  

•    os.system("unzip -u cifar10.zip")
  

•    os.system("mv cifar/\* .; rm -rf cifar; rm cifar10.zip")
  

•    url = 'http://webdocs.cs.ualberta.ca/~bx3/data/cifar10.zip'
  

•    if sys.version_info.major == 2:
  

•        from urllib import urlretrieve
  

•    elif sys.version_info.major == 3:
  

•        from urllib.request import urlretrieve
  

•    urlretrieve(url, 'cifar10.zip')
  

•    import zipfile
  

Yes, zipfile is in python standard library.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/dmlc/mxnet/pull/2325/files/ce5b4b9f8f789db8cb6799de341726d60ce3caa1#r65758312,
or mute the thread
https://github.com/notifications/unsubscribe/ABFI4WfsBu9cFwBt8TuOMteeMFXLep6wks5qIHhBgaJpZM4ItTgY
.

HONG Chuntao
System Research Group
Microsoft Research Asia

[piiswrong]

@hjk41 Could you verify thsi?

[piiswrong]

LGTM. Please rebase and merge

[piiswrong]

There seems to be a lot of new changes. Could you explain them?

@thirdwing Seems to be about R package

[thirdwing]

It is from #2329

All the commits in that PR have been squashed into one commit.

[CharlesMei]

I'm sorry, i'm not familiar with git to do this. I think i may need to learn more about git. Can you help me?

[piiswrong]

try
git origin add upstream https://github.com/dmlc/mxnet.git
git fetch upstream
git rebase upstream/master

[tqchen]

another simpler way might be simply keep your changes and start from a clean fork

[CharlesMei]

@piiswrong, thank you , but because of my former mistakes, my master has changed, i'm afraid there may be something wrong to just do this now.

[CharlesMei]

@tqchen do you mean remove my master branch, and rebase my dev on upstream

[sxjscience]

@CharlesMei Try the method here http://eli.thegreenplace.net/2014/02/19/squashing-github-pull-requests-into-a-single-commit . You can, for example, use git rebase -i dmlc/master

[sxjscience]

@CharlesMei You need to update the submodules.