Add support for tar.gz archive files to download from download_from_google_storage

download_from_google_storage.py

Index: download_from_google_storage.py
diff --git a/download_from_google_storage.py b/download_from_google_storage.py
old mode 100755
new mode 100644
index a45f387ce2aeb3a2d9f8f820206ff8e597eb1e5c..3bba49498acae0e3920ebd23f29d869afc7b9cdd
--- a/download_from_google_storage.py
+++ b/download_from_google_storage.py
@@ -11,8 +11,10 @@ import optparse
 import os
 import Queue
 import re
+import shutil
 import stat
 import sys
+import tarfile
 import threading
 import time
 
@@ -49,7 +51,6 @@ def GetNormalizedPlatform():
     return 'win32'
   return sys.platform
 
-
 # Common utilities
 class Gsutil(object):
   """Call gsutil with some predefined settings.  This is a convenience object,
@@ -186,8 +187,19 @@ def enumerate_work_queue(input_filename, work_queue, directory,
   return work_queue_size
 
 
+def _validate_tar_file(tar, prefix):
+  def _validate(tarinfo):
+    """Returns false if the tarinfo is something we explicitly forbid."""
+    if tarinfo.issym() or tarinfo.islnk():
+      return False
+    if '..' in tarinfo.name or not tarinfo.name.startswith(prefix):
+      return False
+    return True
+  return all(map(_validate, tar.getmembers()))
+
 def _downloader_worker_thread(thread_num, q, force, base_url,
-                              gsutil, out_q, ret_codes, verbose):
+                              gsutil, out_q, ret_codes, verbose, extract,
+                              delete=True):
   while True:
     input_sha1_sum, output_filename = q.get()
     if input_sha1_sum is None:
@@ -218,7 +230,8 @@ def _downloader_worker_thread(thread_num, q, force, base_url,
     # Fetch the file.
     out_q.put('%d> Downloading %s...' % (thread_num, output_filename))
     try:
-      os.remove(output_filename)  # Delete the file if it exists already.
+      if delete:
+        os.remove(output_filename)  # Delete the file if it exists already.
     except OSError:
       if os.path.exists(output_filename):
         out_q.put('%d> Warning: deleting %s failed.' % (
@@ -228,6 +241,34 @@ def _downloader_worker_thread(thread_num, q, force, base_url,
       out_q.put('%d> %s' % (thread_num, err))
       ret_codes.put((code, err))
 
+    if extract:
+      if (not tarfile.is_tarfile(output_filename)
+          or not output_filename.endswith('.tar.gz')):
+        out_q.put('%d> Error: %s is not a tar.gz archive.' % (
+                  thread_num, output_filename))
+        ret_codes.put((1, '%s is not a tar.gz archive.' % (output_filename)))
+        continue
+      tar = tarfile.open(output_filename, 'r:gz')

[hinoka, 2015/06/24 18:54:38]

with tarfile.open(...) as tar:

[ricow1, 2015/06/25 06:24:05]

Done.

+      dirname = os.path.dirname(os.path.abspath(output_filename))
+      extract_dir = output_filename[0:len(output_filename)-7]
+      if not _validate_tar_file(tar, os.path.basename(extract_dir)):
+        out_q.put('%d> Error: %s contains files outside %s.' % (
+                  thread_num, output_filename, extract_dir))
+        ret_codes.put((1, '%s contains invalid entries.' % (output_filename)))
+        continue
+      out_q.put('%d> Extracting %s...' % (thread_num, extract_dir))
+      if os.path.exists(extract_dir):
+        try:
+          shutil.rmtree(extract_dir)
+          out_q.put('%d> Removed %s...' % (thread_num, extract_dir))
+        except OSError:
+          out_q.put('%d> Warning: Can\'t delete: %s' % (
+                    thread_num, extract_dir))
+          ret_codes.put((1, 'Can\'t delete %s.' % (extract_dir)))
+          continue
+        out_q.put('%d> Extracting %s to %s' % (thread_num, output_filename,

[hinoka, 2015/06/24 18:54:38]

Is this supposed to be in this if block? Also seems redundant with line 259.

Lets also add in # of files being extracted to the message.

[ricow1, 2015/06/25 06:24:04]

Removed printing above, added count

+                                               extract_dir))
+      tar.extractall(path=dirname)
     # Set executable bit.
     if sys.platform == 'cygwin':
       # Under cygwin, mark all files as executable. The executable flag in
@@ -258,7 +299,7 @@ def printer_worker(output_queue):
 
 def download_from_google_storage(
     input_filename, base_url, gsutil, num_threads, directory, recursive,
-    force, output, ignore_errors, sha1_file, verbose, auto_platform):
+    force, output, ignore_errors, sha1_file, verbose, auto_platform, extract):
   # Start up all the worker threads.
   all_threads = []
   download_start = time.time()
@@ -270,7 +311,7 @@ def download_from_google_storage(
     t = threading.Thread(
         target=_downloader_worker_thread,
         args=[thread_num, work_queue, force, base_url,
-              gsutil, stdout_queue, ret_codes, verbose])
+              gsutil, stdout_queue, ret_codes, verbose, extract])
     t.daemon = True
     t.start()
     all_threads.append(t)
@@ -358,6 +399,13 @@ def main(args):
                          '(linux|mac|win).  If so, the script will only '
                          'process files that are in the paths that '
                          'that matches the current platform.')
+  parser.add_option('-u', '--extract',
+                    action='store_true',
+                    help='Extract a downloaded tar.gz file. '
+                         'Leaves the tar.gz file around for sha1 verification'
+                         'If a directory with the same name as the tar.gz '
+                         'file already exists, is deleted (to get a '
+                         'clean state in case of update.)')
   parser.add_option('-v', '--verbose', action='store_true',
                     help='Output extra diagnostic and progress information.')
 
@@ -451,7 +499,8 @@ def main(args):
   return download_from_google_storage(
       input_filename, base_url, gsutil, options.num_threads, options.directory,
       options.recursive, options.force, options.output, options.ignore_errors,
-      options.sha1_file, options.verbose, options.auto_platform)
+      options.sha1_file, options.verbose, options.auto_platform,
+      options.extract)
 
 
 if __name__ == '__main__':