winfreed.py: handle urllib2.urlopen failures better

[winfreed.git] / winfreed.py

diff --git a/winfreed.py b/winfreed.py
index a1639c3..a59d78e 100755 (executable)
--- a/winfreed.py
+++ b/winfreed.py
@@ -63,33 +63,42 @@ def show(pkg):
     print
 
 
-def download(pkg):
-    # the "%s" in URLs are meant to be replaced with LANGCODE
+def download_file(src_url, dest_dir):
     try:
-        url = pkg['URL'] % LANGCODE
-    except:
-        url = pkg['URL']
-        pass
+        response = urllib2.urlopen(src_url)
+    except Exception, e:
+        if hasattr(e, 'reason'):
+            print 'Download failed. Reason: ', e.reason
+        elif hasattr(e, 'code'):
+            print 'Download failed. Error code: ', e.code
+        return
 
-    response = urllib2.urlopen(url)
+    # get the final URL after possible redirect have been followed
+    url = response.geturl()
 
     filename = ""
-    if 'Content-Disposition' in  response.info():
-        # Use the filename the server tells us if any,
-        # re pattern from http://stackoverflow.com/questions/8035900
-        content_disposition = response.info().getheader('Content-Disposition').strip()
-        filename = re.findall("filename=(\S+)", content_disposition)[0]
-    
-    if filename == "":
-        filename = urllib2.unquote(os.path.basename(response.geturl()))
+
+    # From http://paste.pound-python.org/show/9545/
+    # TODO: use a proper module to parse HTTP headers
+    if 'Content-Disposition' in response.info() and len(response.info()['Content-Disposition'].split('filename=')) > 1:
+        # If the response has Content-Disposition, we take file name from it
+        filename = response.info()['Content-Disposition'].split('filename=')[1].decode('utf-8')
+        if filename[0] == '"' or filename[0] == "'":
+            filename = urllib2.unquote(filename.split('"')[1])
+    else:
+        filename = urllib2.unquote(url.split('/')[-1].decode('utf_8'))
 
     if filename == "":
-        sys.stderr.write("Debug (%s): filename: %s url: %s\n" %(pkg['package_name'], filename, response.geturl()))
+        sys.stderr.write("Debug (%s): filename: %s url: %s\n" % (pkg['package_name'], filename, url))
         return
 
-    destfile = os.path.join(OUTPUT_DIR, filename)
+    # TODO: Add some integrity verification of downloaded files (md5, sha256?)
+
+    destfile = os.path.join(dest_dir, filename)
     if os.path.exists(destfile):
-        sys.stderr.write("Warning (%s): %s exists!\n" % (pkg['package_name'], destfile))
+        # TODO: check if the file is a full download from previous run,
+        # if not download again discarding the existing file?
+        sys.stderr.write("Warning: %s exists!\n" % destfile)
         return
 
     outfile = open(destfile, mode='w')
@@ -97,7 +106,8 @@ def download(pkg):
     total_size = response.info().getheader('Content-Length').strip()
     total_size = int(total_size)
 
-    widgets = [pkg['name'], ' ', Percentage(), ' ', Bar(marker='=', left='[', right=']'),
+    #widgets = [pkg['name'], ' ', Percentage(), ' ', Bar(marker='=', left='[', right=']'),
+    widgets = [filename, ' ', Percentage(), ' ', Bar(marker='=', left='[', right=']'),
                ' ', ETA(), ' ', FileTransferSpeed()]
     pbar = ProgressBar(widgets=widgets, maxval=total_size).start()
 
@@ -112,6 +122,19 @@ def download(pkg):
         pbar.update(bytes_so_far)
     pbar.finish()
 
+    outfile.close()
+
+
+def download(pkg):
+    # the "%s" in URLs are meant to be replaced with LANGCODE
+    try:
+        pkg_url = pkg['URL'] % LANGCODE
+    except:
+        pkg_url = pkg['URL']
+        pass
+
+    download_file(pkg_url, OUTPUT_DIR)
+
 
 def show_all():
     process_all(PKG_DIR, show)
@@ -145,5 +168,5 @@ if __name__ == "__main__":
     else:
         usage()
         sys.exit(1)
-    
+
     sys.exit(0)