在 Python 中使用 mnist 库

Question

我需要使用mnist Python 库来下载和读取 MNIST 数据（PyPi、Github）：

import mnist

mnist_dataset = mnist.test_images().astype(np.float32)

在我的大学集群上，我毫无问题地加载数据。但是，在我的 PC 上本地，我得到：

---------------------------------------------------------------------------
HTTPError                                 Traceback (most recent call last)
<ipython-input-24-78f59b728818> in <module>
      8 DATASET_SIZE = 512
      9 DIGIT_SIZE = 28
---> 10 mnist_dataset = mnist.test_images().astype(np.float32)
     11 np.random.shuffle(mnist_dataset)
     12 mnist_dataset = np.reshape(mnist_dataset[:DATASET_SIZE] / 255.0, newshape=(DATASET_SIZE, DIGIT_SIZE*DIGIT_SIZE))

~\anaconda3\lib\site-packages\mnist\__init__.py in test_images()
    174         columns of the image
    175     """
--> 176     return download_and_parse_mnist_file('t10k-images-idx3-ubyte.gz')
    177 
    178 

~\anaconda3\lib\site-packages\mnist\__init__.py in download_and_parse_mnist_file(fname, target_dir, force)
    141         Numpy array with the dimensions and the data in the IDX file
    142     """
--> 143     fname = download_file(fname, target_dir=target_dir, force=force)
    144     fopen = gzip.open if os.path.splitext(fname)[1] == '.gz' else open
    145     with fopen(fname, 'rb') as fd:

~\anaconda3\lib\site-packages\mnist\__init__.py in download_file(fname, target_dir, force)
     57     if force or not os.path.isfile(target_fname):
     58         url = urljoin(datasets_url, fname)
---> 59         urlretrieve(url, target_fname)
     60 
     61     return target_fname

~\anaconda3\lib\urllib\request.py in urlretrieve(url, filename, reporthook, data)
    245     url_type, path = splittype(url)
    246 
--> 247     with contextlib.closing(urlopen(url, data)) as fp:
    248         headers = fp.info()
    249 

~\anaconda3\lib\urllib\request.py in urlopen(url, data, timeout, cafile, capath, cadefault, context)
    220     else:
    221         opener = _opener
--> 222     return opener.open(url, data, timeout)
    223 
    224 def install_opener(opener):

~\anaconda3\lib\urllib\request.py in open(self, fullurl, data, timeout)
    529         for processor in self.process_response.get(protocol, []):
    530             meth = getattr(processor, meth_name)
--> 531             response = meth(req, response)
    532 
    533         return response

~\anaconda3\lib\urllib\request.py in http_response(self, request, response)
    639         if not (200 <= code < 300):
    640             response = self.parent.error(
--> 641                 'http', request, response, code, msg, hdrs)
    642 
    643         return response

~\anaconda3\lib\urllib\request.py in error(self, proto, *args)
    567         if http_err:
    568             args = (dict, 'default', 'http_error_default') + orig_args
--> 569             return self._call_chain(*args)
    570 
    571 # XXX probably also want an abstract factory that knows when it makes

~\anaconda3\lib\urllib\request.py in _call_chain(self, chain, kind, meth_name, *args)
    501         for handler in handlers:
    502             func = getattr(handler, meth_name)
--> 503             result = func(*args)
    504             if result is not None:
    505                 return result

~\anaconda3\lib\urllib\request.py in http_error_default(self, req, fp, code, msg, hdrs)
    647 class HTTPDefaultErrorHandler(BaseHandler):
    648     def http_error_default(self, req, fp, code, msg, hdrs):
--> 649         raise HTTPError(req.full_url, code, msg, hdrs, fp)
    650 
    651 class HTTPRedirectHandler(BaseHandler):

HTTPError: HTTP Error 403: Forbidden

我已经通过inspect 模块检查了这些功能。本地版本和集群版本调用的HTTP地址相同，即http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz。我可以从我的网络浏览器毫无问题地下载它。

我可以在 Python 中对此做些什么？

Answer 1

这个模块已经很老了，已经存档了。

我预计它可能会使用具有新安全系统的新服务器，并且代码可能需要一些设置（例如标头 User-Agent）才能正确访问数据。

---

使用来自@wwii 评论的建议，我下载了源代码并添加了User-Agent
现在我可以下载图片了

mint/__init__.py

try:
    #from urllib.request import urlretrieve            # before
    from urllib.request import urlretrieve, URLopener  # after
except ImportError:
    #from urllib import urlretrieve # py2            # before
    from urllib import urlretrieve, URLopener # py2  # after


# ... code ...


def download_file(fname, target_dir=None, force=False):

    # ... code ...

    if force or not os.path.isfile(target_fname):
        url = urljoin(datasets_url, fname)

        # before
        #urlretrieve(url, target_fname)   

        # after
        opener = URLopener()
        opener.addheader('User-Agent', "Mozilla/5.0")
        opener.retrieve(url, target_fname)

---

测试代码：

import mnist
import numpy as np

print(mnist.__file__)         # to see if I uses local version with changes
print(mnist.datasets_url)     
print(mnist.temporary_dir())  # to see where it is downloaded

mnist_dataset = mnist.test_images().astype(np.float32)
print(mnist_dataset)

---

仅在 Python 3.8 和 Python 2.x 上测试过

Answer 2

另一种解决方案是自己滚动来下载文件。使用我找到的函数 here 和来自 mnist site 的文件 url ...

import  requests, gzip

urls = [(r'http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz','training_images.gz'),
        (r'http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz','training_labels.gz'),
        (r'http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz','test_images.gz'),
        (r'http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz','test_labels.gz')]

def download_url(url, save_path, chunk_size=128):
    r = requests.get(url, stream=True)
    with open(save_path, 'wb') as fd:
        for chunk in r.iter_content(chunk_size=chunk_size):
            fd.write(chunk)
for url,path in urls:
    download_url(url,path)

---

这很好，它只是将 zip 文件下载到当前工作目录。你仍然需要unzip them。

Answer 3

使用其他答案，我已经能够构建一个允许直接使用包的解决方案。

以下代码必须执行一次并在全局范围内工作：

from six.moves import urllib

opener = urllib.request.build_opener()
opener.addheaders = [('User-agent', 'Mozilla/5.0')]
urllib.request.install_opener(opener)

这样，所有mnist 功能都可以使用，文件将根据需要下载。在调用 mnist_dataset = mnist.test_images().astype(np.float32) 之前，我直接在单元格中的 Jupyter Notebook 中执行上面的代码。