Handle unknown media length (#406)

Some media download APIs does not support partial downloads and do not return a 'Content-length' or 'Content-range' header and is therefore never considered done. This change is to consider responses where the media length is unknown to be done. This commit could potentially break media downloads where the API supports partial download and only returns a 'Content-length' header when the entire file has been served/read. This commit fixes issue #15

Author: danielx

docs/epy/googleapiclient.http-pysrc.html

@@ -844,7 +844,7 @@ <h1 class="epydoc">Source Code for <a href="googleapiclient.http-module.html">Mo
 <a name="L655"></a><tt class="py-lineno"> 655</tt> <tt class="py-line"><tt class="py-docstring"> Returns:</tt> </tt>
 <a name="L656"></a><tt class="py-lineno"> 656</tt> <tt class="py-line"><tt class="py-docstring"> (status, done): (MediaDownloadStatus, boolean)</tt> </tt>
 <a name="L657"></a><tt class="py-lineno"> 657</tt> <tt class="py-line"><tt class="py-docstring"> The value of 'done' will be True when the media has been fully</tt> </tt>
-<a name="L658"></a><tt class="py-lineno"> 658</tt> <tt class="py-line"><tt class="py-docstring"> downloaded.</tt> </tt>
+<a name="L658"></a><tt class="py-lineno"> 658</tt> <tt class="py-line"><tt class="py-docstring"> downloaded or the total size of the media is unknown.</tt> </tt>
 <a name="L659"></a><tt class="py-lineno"> 659</tt> <tt class="py-line"><tt class="py-docstring"></tt> </tt>
 <a name="L660"></a><tt class="py-lineno"> 660</tt> <tt class="py-line"><tt class="py-docstring"> Raises:</tt> </tt>
 <a name="L661"></a><tt class="py-lineno"> 661</tt> <tt class="py-line"><tt class="py-docstring"> googleapiclient.errors.HttpError if the response was not a 2xx.</tt> </tt>
@@ -873,7 +873,7 @@ <h1 class="epydoc">Source Code for <a href="googleapiclient.http-module.html">Mo
 <a name="L684"></a><tt class="py-lineno"> 684</tt> <tt class="py-line"> <tt class="py-keyword">elif</tt> <tt class="py-string">'content-length'</tt> <tt class="py-keyword">in</tt> <tt class="py-name">resp</tt><tt class="py-op">:</tt> </tt>
 <a name="L685"></a><tt class="py-lineno"> 685</tt> <tt class="py-line"> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_total_size</tt> <tt class="py-op">=</tt> <tt class="py-name">int</tt><tt class="py-op">(</tt><tt class="py-name">resp</tt><tt class="py-op">[</tt><tt class="py-string">'content-length'</tt><tt class="py-op">]</tt><tt class="py-op">)</tt> </tt>
 <a name="L686"></a><tt class="py-lineno"> 686</tt> <tt class="py-line"> </tt>
-<a name="L687"></a><tt class="py-lineno"> 687</tt> <tt class="py-line"> <tt class="py-keyword">if</tt> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_progress</tt> <tt class="py-op">==</tt> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_total_size</tt><tt class="py-op">:</tt> </tt>
+<a name="L687"></a><tt class="py-lineno"> 687</tt> <tt class="py-line"> <tt class="py-keyword">if</tt> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_total_size</tt> <tt class="py-keyword">is</tt> <tt class="py-name">None</tt> <tt class="py-keyword">or</tt> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_progress</tt> <tt class="py-op">==</tt> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_total_size</tt><tt class="py-op">:</tt> </tt>
 <a name="L688"></a><tt class="py-lineno"> 688</tt> <tt class="py-line"> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_done</tt> <tt class="py-op">=</tt> <tt class="py-name">True</tt> </tt>
 <a name="L689"></a><tt class="py-lineno"> 689</tt> <tt class="py-line"> <tt class="py-keyword">return</tt> <tt id="link-86" class="py-name" targets="Class googleapiclient.http.MediaDownloadProgress=googleapiclient.http.MediaDownloadProgress-class.html"><a title="googleapiclient.http.MediaDownloadProgress" class="py-name" href="#" onclick="return doclink('link-86', 'MediaDownloadProgress', 'link-86');">MediaDownloadProgress</a></tt><tt class="py-op">(</tt><tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_progress</tt><tt class="py-op">,</tt> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_total_size</tt><tt class="py-op">)</tt><tt class="py-op">,</tt> <tt class="py-name">self</tt><tt class="py-op">.</tt><tt class="py-name">_done</tt> </tt>
 <a name="L690"></a><tt class="py-lineno"> 690</tt> <tt class="py-line"> <tt class="py-keyword">else</tt><tt class="py-op">:</tt> </tt>
@@ -2189,7 +2189,7 @@ <h1 class="epydoc">Source Code for <a href="googleapiclient.http-module.html">Mo
 <table border="0" cellpadding="0" cellspacing="0" width="100%%">
  <tr>
  <td align="left" class="footer">
- Generated by Epydoc 3.0.1 on Mon Jun  5 13:38:40 2017
+ Generated by Epydoc 3.0.1 on Wed Jun 14 20:18:21 2017
  </td>
  <td align="right" class="footer">
  <a target="mainFrame" href="http://epydoc.sourceforge.net"

docs/epy/googleapiclient.http.MediaIoBaseDownload-class.html

@@ -266,7 +266,7 @@ <h3 class="epydoc"><span class="sig"><span class="sig-name">next_chunk</span>(<s
 Returns:
  (status, done): (MediaDownloadStatus, boolean)
  The value of 'done' will be True when the media has been fully
- downloaded.
+ downloaded or the total size of the media is unknown.
 
 Raises:
  googleapiclient.errors.HttpError if the response was not a 2xx.
@@ -308,7 +308,7 @@ <h3 class="epydoc"><span class="sig"><span class="sig-name">next_chunk</span>(<s
 <table border="0" cellpadding="0" cellspacing="0" width="100%%">
  <tr>
  <td align="left" class="footer">
- Generated by Epydoc 3.0.1 on Mon Jun  5 13:38:40 2017
+ Generated by Epydoc 3.0.1 on Wed Jun 14 20:18:20 2017
  </td>
  <td align="right" class="footer">
  <a target="mainFrame" href="http://epydoc.sourceforge.net"

googleapiclient/http.py

@@ -658,7 +658,7 @@ def next_chunk(self, num_retries=0):
  Returns:
  (status, done): (MediaDownloadProgress, boolean)
  The value of 'done' will be True when the media has been fully
- downloaded.
+ downloaded or the total size of the media is unknown.
 
  Raises:
  googleapiclient.errors.HttpError if the response was not a 2xx.
@@ -687,7 +687,7 @@ def next_chunk(self, num_retries=0):
  elif 'content-length' in resp:
  self._total_size = int(resp['content-length'])
 
- if self._progress == self._total_size:
+ if self._total_size is None or self._progress == self._total_size:
  self._done = True
  return MediaDownloadProgress(self._progress, self._total_size), self._done
  else:

tests/test_http.py

@@ -576,6 +576,29 @@ def test_media_io_base_download_empty_file(self):
  self.assertEqual(0, download._total_size)
  self.assertEqual(0, status.progress())
 
+ def test_media_io_base_download_unknown_media_size(self):
+ self.request.http = HttpMockSequence([
+ ({'status': '200'}, b'123')
+ ])
+
+ download = MediaIoBaseDownload(
+ fd=self.fd, request=self.request, chunksize=3)
+
+ self.assertEqual(self.fd, download._fd)
+ self.assertEqual(0, download._progress)
+ self.assertEqual(None, download._total_size)
+ self.assertEqual(False, download._done)
+ self.assertEqual(self.request.uri, download._uri)
+
+ status, done = download.next_chunk()
+
+ self.assertEqual(self.fd.getvalue(), b'123')
+ self.assertEqual(True, done)
+ self.assertEqual(3, download._progress)
+ self.assertEqual(None, download._total_size)
+ self.assertEqual(0, status.progress())
+
+
 EXPECTED = """POST /someapi/v1/collection/?foo=bar HTTP/1.1
 Content-Type: application/json
 MIME-Version: 1.0