googlecloud 0.0.4 → 0.0.6

data/packages/gsutil/gslib/tests/util.py

@@ -0,0 +1,125 @@
+# Copyright 2013 Google Inc. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import posixpath
+import os.path
+import time
+import urllib
+import urlparse
+
+import unittest
+if not hasattr(unittest.TestCase, 'assertIsNone'):
+  # external dependency unittest2 required for Python <= 2.6
+  import unittest2 as unittest
+
+# Flags for running different types of tests.
+RUN_INTEGRATION_TESTS = True
+RUN_UNIT_TESTS = True
+
+# Whether the tests are running verbose or not.
+VERBOSE_OUTPUT = False
+
+
+def _NormalizeURI(uri):
+  """Normalizes the path component of a URI.
+
+  Examples:
+    gs://foo//bar -> gs://foo/bar
+    gs://foo/./bar -> gs://foo/bar
+  """
+  # Note: we have to do this dance of changing gs:// to file:// because on
+  # Windows, the urlparse function won't work with URL schemes that are not
+  # known. urlparse('gs://foo/bar') on Windows turns into:
+  #     scheme='gs', netloc='', path='//foo/bar'
+  # while on non-Windows platforms, it turns into:
+  #     scheme='gs', netloc='foo', path='/bar'
+  uri = uri.replace('gs://', 'file://')
+  parsed = list(urlparse.urlparse(uri))
+  parsed[2] = posixpath.normpath(parsed[2])
+  if parsed[2].startswith('//'):
+    # The normpath function doesn't change '//foo' -> '/foo' by design.
+    parsed[2] = parsed[2][1:]
+  unparsed = urlparse.urlunparse(parsed)
+  unparsed = unparsed.replace('file://', 'gs://')
+  return unparsed
+
+
+def ObjectToURI(obj, *suffixes):
+  """Returns the storage URI string for a given StorageUri or file object.
+
+  Args:
+    obj: The object to get the URI from. Can be a file object, a subclass of
+         boto.storage_uri.StorageURI, or a string. If a string, it is assumed to
+         be a local on-disk path.
+    suffixes: Suffixes to append. For example, ObjectToUri(bucketuri, 'foo')
+              would return the URI for a key name 'foo' inside the given bucket.
+  """
+  if isinstance(obj, file):
+    return 'file://%s' % os.path.abspath(os.path.join(obj.name, *suffixes))
+  if isinstance(obj, basestring):
+    return 'file://%s' % os.path.join(obj, *suffixes)
+  uri = obj.uri
+  if suffixes:
+    uri = _NormalizeURI('/'.join([uri] + list(suffixes)))
+
+  # Storage URIs shouldn't contain a trailing slash.
+  if uri.endswith('/'):
+    uri = uri[:-1]
+  return uri
+
+
+def Retry(ExceptionToCheck, tries=4, delay=3, backoff=2, logger=None):
+  """Retry calling the decorated function using an exponential backoff.
+
+  Taken from:
+    https://github.com/saltycrane/retry-decorator
+  Licensed under BSD:
+    https://github.com/saltycrane/retry-decorator/blob/master/LICENSE
+
+  :param ExceptionToCheck: the exception to check. may be a tuple of
+      exceptions to check
+  :type ExceptionToCheck: Exception or tuple
+  :param tries: number of times to try (not retry) before giving up
+  :type tries: int
+  :param delay: initial delay between retries in seconds
+  :type delay: int
+  :param backoff: backoff multiplier e.g. value of 2 will double the delay
+      each retry
+  :type backoff: int
+  :param logger: logger to use. If None, print
+  :type logger: logging.Logger instance
+  """
+  def deco_retry(f):
+    def f_retry(*args, **kwargs):
+      mtries, mdelay = tries, delay
+      try_one_last_time = True
+      while mtries > 1:
+        try:
+          return f(*args, **kwargs)
+          try_one_last_time = False
+          break
+        except ExceptionToCheck, e:
+          msg = "%s, Retrying in %d seconds..." % (str(e), mdelay)
+          if logger:
+              logger.warning(msg)
+          else:
+              print msg
+          time.sleep(mdelay)
+          mtries -= 1
+          mdelay *= backoff
+      if try_one_last_time:
+        return f(*args, **kwargs)
+      return
+    return f_retry  # true decorator
+  return deco_retry

data/packages/gsutil/gslib/thread_pool.py

@@ -0,0 +1,79 @@
+# Copyright 2011 Google Inc. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Basic thread pool with exception handler."""
+
+import logging
+import Queue
+import threading
+
+
+# Magic values used to cleanly bring down threads.
+_THREAD_EXIT_MAGIC = ('Clean', 'Thread', 'Exit')
+
+
+def _DefaultExceptionHandler(e):
+  logging.exception(e)
+
+
+class Worker(threading.Thread):
+  """Thread executing tasks from a given task's queue."""
+
+  def __init__(self, tasks, exception_handler):
+    threading.Thread.__init__(self)
+    self.tasks = tasks
+    self.daemon = True
+    self.exception_handler = exception_handler
+    self.start()
+
+  def run(self):
+    while True:
+      func, args, kargs = self.tasks.get()
+
+      # Listen for magic value indicating thread exit.
+      if (func, args, kargs) == _THREAD_EXIT_MAGIC:
+        break
+
+      try:
+        func(*args, **kargs)
+      except Exception, e:
+        self.exception_handler(e)
+      finally:
+        self.tasks.task_done()
+
+
+class ThreadPool(object):
+  """Pool of threads consuming tasks from a queue."""
+
+  def __init__(self, num_threads, exception_handler=_DefaultExceptionHandler):
+    self.tasks = Queue.Queue(num_threads)
+    self.threads = []
+    for _ in range(num_threads):
+      self.threads.append(Worker(self.tasks, exception_handler))
+
+  def AddTask(self, func, *args, **kargs):
+    """Add a task to the queue."""
+    self.tasks.put((func, args, kargs))
+
+  def WaitCompletion(self):
+    """Wait for completion of all the tasks in the queue."""
+    self.tasks.join()
+
+  def Shutdown(self):
+    """Shutdown the thread pool."""
+    for thread in self.threads:
+      self.tasks.put(_THREAD_EXIT_MAGIC)
+
+    for thread in self.threads:
+      thread.join()

data/packages/gsutil/gslib/util.py

@@ -0,0 +1,151 @@
+# Copyright 2010 Google Inc. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Static data and helper functions."""
+
+import math
+import re
+import sys
+
+import boto
+
+# We don't use the oauth2 authentication plugin directly; importing it here
+# ensures that it's loaded and available by default. Note: we made this static
+# state instead of Command instance state because the top-level gsutil code
+# needs to check it.
+HAVE_OAUTH2 = False
+try:
+  from oauth2_plugin import oauth2_helper
+  HAVE_OAUTH2 = True
+except ImportError:
+  pass
+
+TWO_MB = 2 * 1024 * 1024
+
+NO_MAX = sys.maxint
+
+# Binary exponentiation strings.
+_EXP_STRINGS = [
+  (0, 'B', 'bit'),
+  (10, 'KB', 'kbit'),
+  (20, 'MB', 'Mbit'),
+  (30, 'GB', 'Gbit'),
+  (40, 'TB', 'Tbit'),
+  (50, 'PB', 'Pbit'),
+]
+
+# Detect platform types.
+IS_WINDOWS = 'win32' in str(sys.platform).lower()
+IS_LINUX = 'linux' in str(sys.platform).lower()
+IS_OSX = 'darwin' in str(sys.platform).lower()
+
+# Enum class for specifying listing style.
+class ListingStyle(object):
+  SHORT = 'SHORT'
+  LONG = 'LONG'
+  LONG_LONG = 'LONG_LONG'
+
+
+def HasConfiguredCredentials():
+  """Determines if boto credential/config file exists."""
+  config = boto.config
+  has_goog_creds = (config.has_option('Credentials', 'gs_access_key_id') and
+                    config.has_option('Credentials', 'gs_secret_access_key'))
+  has_amzn_creds = (config.has_option('Credentials', 'aws_access_key_id') and
+                    config.has_option('Credentials', 'aws_secret_access_key'))
+  has_oauth_creds = (HAVE_OAUTH2 and
+      config.has_option('Credentials', 'gs_oauth2_refresh_token'))
+  has_auth_plugins = config.has_option('Plugin', 'plugin_directory')
+  return (has_goog_creds or has_amzn_creds or has_oauth_creds
+          or has_auth_plugins)
+
+
+def _RoundToNearestExponent(num):
+  i = 0
+  while i+1 < len(_EXP_STRINGS) and num >= (2 ** _EXP_STRINGS[i+1][0]):
+    i += 1
+  return i, round(float(num) / 2 ** _EXP_STRINGS[i][0], 2)
+
+def MakeHumanReadable(num):
+  """Generates human readable string for a number of bytes.
+
+  Args:
+    num: The number, in bytes.
+
+  Returns:
+    A string form of the number using size abbreviations (KB, MB, etc.).
+  """
+  i, rounded_val = _RoundToNearestExponent(num)
+  return '%s %s' % (rounded_val, _EXP_STRINGS[i][1])
+
+def MakeBitsHumanReadable(num):
+  """Generates human readable string for a number of bits.
+
+  Args:
+    num: The number, in bits.
+
+  Returns:
+    A string form of the number using bit size abbreviations (kbit, Mbit, etc.)
+  """
+  i, rounded_val = _RoundToNearestExponent(num)
+  return '%s %s' % (rounded_val, _EXP_STRINGS[i][2])
+
+def Percentile(values, percent, key=lambda x:x):
+  """Find the percentile of a list of values.
+
+  Taken from: http://code.activestate.com/recipes/511478/
+
+  Args:
+    values: a list of numeric values. Note that the values MUST BE already
+            sorted.
+    percent: a float value from 0.0 to 1.0.
+    key: optional key function to compute value from each element of the list
+         of values.
+
+  Returns:
+    The percentile of the values.
+  """
+  if not values:
+    return None
+  k = (len(values) - 1) * percent
+  f = math.floor(k)
+  c = math.ceil(k)
+  if f == c:
+    return key(values[int(k)])
+  d0 = key(values[int(f)]) * (c-k)
+  d1 = key(values[int(c)]) * (k-f)
+  return d0 + d1
+
+def ExtractErrorDetail(e):
+  """Extract <Details> text from XML content.
+
+  Args:
+    e: The GSResponseError that includes XML to be parsed.
+
+  Returns:
+    (exception_name, d), where d is <Details> text or None if not found.
+  """
+  exc_name_parts = re.split("[\.']", str(type(e)))
+  if len(exc_name_parts) < 2:
+    # Shouldn't happen, but have fallback in case.
+    exc_name = str(type(e))
+  else:
+    exc_name = exc_name_parts[-2]
+  if not hasattr(e, 'body'):
+    return (exc_name, None)
+  detail_start = e.body.find('<Details>')
+  detail_end = e.body.find('</Details>')
+  if detail_start != -1 and detail_end != -1:
+    return (exc_name, e.body[detail_start+9:detail_end])
+  return (exc_name, None)

data/packages/gsutil/gslib/wildcard_iterator.py

@@ -0,0 +1,492 @@
+# Copyright 2010 Google Inc. All Rights Reserved.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish, dis-
+# tribute, sublicense, and/or sell copies of the Software, and to permit
+# persons to whom the Software is furnished to do so, subject to the fol-
+# lowing conditions:
+#
+# The above copyright notice and this permission notice shall be included
+# in all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+# OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABIL-
+# ITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT
+# SHALL THE AUTHOR BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+# WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+# IN THE SOFTWARE.
+
+"""Implementation of wildcarding over StorageUris.
+
+StorageUri is an abstraction that Google introduced in the boto library,
+for representing storage provider-independent bucket and object names with
+a shorthand URI-like syntax (see boto/boto/storage_uri.py) The current
+class provides wildcarding support for StorageUri objects (including both
+bucket and file system objects), allowing one to express collections of
+objects with syntax like the following:
+  gs://mybucket/images/*.png
+  file:///tmp/???abc???
+
+We provide wildcarding support as part of gsutil rather than as part
+of boto because wildcarding is really part of shell command-like
+functionality.
+
+A comment about wildcard semantics: We support both single path component
+wildcards (e.g., using '*') and recursive wildcards (using '**'), for both
+file and cloud URIs. For example,
+  gs://bucket/doc/*/*.html
+would enumerate HTML files one directory down from gs://bucket/doc, while
+  gs://bucket/**/*.html
+would enumerate HTML files in all objects contained in the bucket.
+
+Note also that if you use file system wildcards it's likely your shell
+interprets the wildcarding before passing the command to gsutil. For example:
+  % gsutil cp /opt/eclipse/*/*.html gs://bucket/eclipse
+would likely be expanded by the shell into the following before running gsutil:
+  % gsutil cp /opt/eclipse/RUNNING.html gs://bucket/eclipse
+
+Note also that most shells don't support '**' wildcarding (I think only
+zsh does). If you want to use '**' wildcarding with such a shell you can
+single quote each wildcarded string, so it gets passed uninterpreted by the
+shell to gsutil (at which point gsutil will perform the wildcarding expansion):
+  % gsutil cp '/opt/eclipse/**/*.html' gs://bucket/eclipse
+"""
+
+import boto
+import fnmatch
+import glob
+import os
+import re
+import sys
+import urllib
+
+from boto.s3.prefix import Prefix
+from boto.storage_uri import BucketStorageUri
+from bucket_listing_ref import BucketListingRef
+
+# Regex to determine if a string contains any wildcards.
+WILDCARD_REGEX = re.compile('[*?\[\]]')
+
+WILDCARD_OBJECT_ITERATOR = 'wildcard_object_iterator'
+WILDCARD_BUCKET_ITERATOR = 'wildcard_bucket_iterator'
+
+
+class WildcardIterator(object):
+  """Base class for wildcarding over StorageUris.
+
+  This class implements support for iterating over StorageUris that
+  contain wildcards.
+
+  The base class is abstract; you should instantiate using the
+  wildcard_iterator() static factory method, which chooses the right
+  implementation depending on the StorageUri.
+  """
+
+  def __repr__(self):
+    """Returns string representation of WildcardIterator."""
+    return 'WildcardIterator(%s)' % self.wildcard_uri
+
+
+class CloudWildcardIterator(WildcardIterator):
+  """WildcardIterator subclass for buckets and objects.
+
+  Iterates over BucketListingRef matching the StorageUri wildcard. It's
+  much more efficient to request the Key from the BucketListingRef (via
+  GetKey()) than to request the StorageUri and then call uri.get_key()
+  to retrieve the key, for cases where you want to get metadata that's
+  available in the Bucket (for example to get the name and size of
+  each object), because that information is available in the bucket GET
+  results. If you were to iterate over URIs for such cases and then get
+  the name and size info from each resulting StorageUri, it would cause
+  an additional object GET request for each of the result URIs.
+  """
+
+  def __init__(self, wildcard_uri, proj_id_handler,
+               bucket_storage_uri_class=BucketStorageUri, all_versions=False,
+               headers=None, debug=0):
+    """
+    Instantiates an iterator over BucketListingRef matching given wildcard URI.
+
+    Args:
+      wildcard_uri: StorageUri that contains the wildcard to iterate.
+      proj_id_handler: ProjectIdHandler to use for current command.
+      bucket_storage_uri_class: BucketStorageUri interface.
+                                Settable for testing/mocking.
+      headers: Dictionary containing optional HTTP headers to pass to boto.
+      debug: Debug level to pass in to boto connection (range 0..3).
+    """
+    self.wildcard_uri = wildcard_uri
+    # Make a copy of the headers so any updates we make during wildcard
+    # expansion aren't left in the input params (specifically, so we don't
+    # include the x-goog-project-id header needed by a subset of cases, in
+    # the data returned to caller, which could then be used in other cases
+    # where that header must not be passed).
+    if headers is None:
+      self.headers = {}
+    else:
+      self.headers = headers.copy()
+    self.proj_id_handler = proj_id_handler
+    self.bucket_storage_uri_class = bucket_storage_uri_class
+    self.all_versions = all_versions
+    self.debug = debug
+
+  def __iter__(self):
+    """Python iterator that gets called when iterating over cloud wildcard.
+
+    Yields:
+      BucketListingRef, or empty iterator if no matches.
+    """
+    # First handle bucket wildcarding, if any.
+    if ContainsWildcard(self.wildcard_uri.bucket_name):
+      regex = fnmatch.translate(self.wildcard_uri.bucket_name)
+      bucket_uris = []
+      prog = re.compile(regex)
+      self.proj_id_handler.FillInProjectHeaderIfNeeded(WILDCARD_BUCKET_ITERATOR,
+                                                       self.wildcard_uri,
+                                                       self.headers)
+      for b in self.wildcard_uri.get_all_buckets(headers=self.headers):
+        if prog.match(b.name):
+          # Use str(b.name) because get_all_buckets() returns Unicode
+          # string, which when used to construct x-goog-copy-src metadata
+          # requests for object-to-object copies causes pathname '/' chars
+          # to be entity-encoded (bucket%2Fdir instead of bucket/dir),
+          # which causes the request to fail.
+          uri_str = '%s://%s' % (self.wildcard_uri.scheme,
+                                 urllib.quote_plus(str(b.name)))
+          bucket_uris.append(
+              boto.storage_uri(
+                  uri_str, debug=self.debug,
+                  bucket_storage_uri_class=self.bucket_storage_uri_class,
+                  suppress_consec_slashes=False))
+    else:
+      bucket_uris = [self.wildcard_uri.clone_replace_name('')]
+
+    # Now iterate over bucket(s), and handle object wildcarding, if any.
+    self.proj_id_handler.FillInProjectHeaderIfNeeded(WILDCARD_OBJECT_ITERATOR,
+                                                     self.wildcard_uri,
+                                                     self.headers)
+    for bucket_uri in bucket_uris:
+      if self.wildcard_uri.names_bucket():
+        # Bucket-only URI.
+        yield BucketListingRef(bucket_uri, key=None, prefix=None,
+                               headers=self.headers)
+      else:
+        # URI contains an object name. If there's no wildcard just yield
+        # the needed URI.
+        if not ContainsWildcard(self.wildcard_uri.object_name):
+          uri_to_yield = bucket_uri.clone_replace_name(
+              self.wildcard_uri.object_name)
+          yield BucketListingRef(uri_to_yield, key=None, prefix=None,
+                          headers=self.headers)
+        else:
+          # URI contains a wildcard. Expand iteratively by building
+          # prefix/delimiter bucket listing request, filtering the results per
+          # the current level's wildcard, and continuing with the next component
+          # of the wildcard. See _BuildBucketFilterStrings() documentation
+          # for details.
+          #
+          # Initialize the iteration with bucket name from bucket_uri but
+          # object name from self.wildcard_uri. This is needed to handle cases
+          # where both the bucket and object names contain wildcards.
+          uris_needing_expansion = [
+              bucket_uri.clone_replace_name(self.wildcard_uri.object_name)]
+          while len(uris_needing_expansion) > 0:
+            uri = uris_needing_expansion.pop(0)
+            (prefix, delimiter, prefix_wildcard, suffix_wildcard) = (
+                self._BuildBucketFilterStrings(uri.object_name))
+            prog = re.compile(fnmatch.translate(prefix_wildcard))
+            # List bucket for objects matching prefix up to delimiter.
+            for key in bucket_uri.list_bucket(prefix=prefix,
+                                              delimiter=delimiter,
+                                              headers=self.headers,
+                                              all_versions=self.all_versions):
+              # Check that the prefix regex matches rstripped key.name (to
+              # correspond with the rstripped prefix_wildcard from
+              # _BuildBucketFilterStrings()).
+              if prog.match(key.name.rstrip('/')):
+                if suffix_wildcard and key.name.rstrip('/') != suffix_wildcard:
+                  if isinstance(key, Prefix):
+                    # There's more wildcard left to expand.
+                    uris_needing_expansion.append(
+                        uri.clone_replace_name(key.name.rstrip('/') + '/'
+                        + suffix_wildcard))
+                else:
+                  # Done expanding.
+                  expanded_uri = uri.clone_replace_key(key)
+
+                  if isinstance(key, Prefix):
+                    yield BucketListingRef(expanded_uri, key=None, prefix=key,
+                                           headers=self.headers)
+                  else:
+                    yield BucketListingRef(expanded_uri, key=key, prefix=None,
+                                           headers=self.headers)
+
+  def _BuildBucketFilterStrings(self, wildcard):
+    """
+    Builds strings needed for querying a bucket and filtering results to
+    implement wildcard object name matching.
+
+    Args:
+      wildcard: The wildcard string to match to objects.
+
+    Returns:
+      (prefix, delimiter, prefix_wildcard, suffix_wildcard)
+      where:
+        prefix is the prefix to be sent in bucket GET request.
+        delimiter is the delimiter to be sent in bucket GET request.
+        prefix_wildcard is the wildcard to be used to filter bucket GET results.
+        suffix_wildcard is wildcard to be appended to filtered bucket GET
+          results for next wildcard expansion iteration.
+      For example, given the wildcard gs://bucket/abc/d*e/f*.txt we
+      would build prefix= abc/d, delimiter=/, prefix_wildcard=d*e, and
+      suffix_wildcard=f*.txt. Using this prefix and delimiter for a bucket
+      listing request will then produce a listing result set that can be
+      filtered using this prefix_wildcard; and we'd use this suffix_wildcard
+      to feed into the next call(s) to _BuildBucketFilterStrings(), for the
+      next iteration of listing/filtering.
+
+    Raises:
+      AssertionError if wildcard doesn't contain any wildcard chars.
+    """
+    # Generate a request prefix if the object name part of the wildcard starts
+    # with a non-wildcard string (e.g., that's true for 'gs://bucket/abc*xyz').
+    match = WILDCARD_REGEX.search(wildcard)
+    if not match:
+      # Input "wildcard" has no wildcard chars, so just return tuple that will
+      # cause a bucket listing to match the given input wildcard. Example: if
+      # previous iteration yielded gs://bucket/dir/ with suffix_wildcard abc,
+      # the next iteration will call _BuildBucketFilterStrings() with
+      # gs://bucket/dir/abc, and we will return prefix ='dir/abc',
+      # delimiter='/', prefix_wildcard='dir/abc', and suffix_wildcard=''.
+      prefix = wildcard
+      delimiter = '/'
+      prefix_wildcard = wildcard
+      suffix_wildcard = ''
+    else:
+      if match.start() > 0:
+        # Wildcard does not occur at beginning of object name, so construct a
+        # prefix string to send to server.
+        prefix = wildcard[:match.start()]
+        wildcard_part = wildcard[match.start():]
+      else:
+        prefix = None
+        wildcard_part = wildcard
+      end = wildcard_part.find('/')
+      if end != -1:
+        wildcard_part = wildcard_part[:end+1]
+      # Remove trailing '/' so we will match gs://bucket/abc* as well as
+      # gs://bucket/abc*/ with the same wildcard regex.
+      prefix_wildcard = ((prefix or '') + wildcard_part).rstrip('/')
+      suffix_wildcard = wildcard[match.end():]
+      end = suffix_wildcard.find('/')
+      if end == -1:
+        suffix_wildcard = ''
+      else:
+        suffix_wildcard = suffix_wildcard[end+1:]
+      # To implement recursive (**) wildcarding, if prefix_wildcard
+      # suffix_wildcard starts with '**' don't send a delimiter, and combine
+      # suffix_wildcard at end of prefix_wildcard.
+      if prefix_wildcard.find('**') != -1:
+        delimiter = None
+        prefix_wildcard = prefix_wildcard + suffix_wildcard
+        suffix_wildcard = ''
+      else:
+        delimiter = '/'
+        delim_pos = suffix_wildcard.find(delimiter)
+    # The following debug output is useful for tracing how the algorithm
+    # walks through a multi-part wildcard like gs://bucket/abc/d*e/f*.txt
+    if self.debug > 1:
+      sys.stderr.write(
+          'DEBUG: wildcard=%s, prefix=%s, delimiter=%s, '
+          'prefix_wildcard=%s, suffix_wildcard=%s\n' %
+          (wildcard, prefix, delimiter, prefix_wildcard, suffix_wildcard))
+    return (prefix, delimiter, prefix_wildcard, suffix_wildcard)
+
+  def IterKeys(self):
+    """
+    Convenience iterator that runs underlying iterator and returns Key for each
+    iteration.
+
+    Yields:
+      Subclass of boto.s3.key.Key, or empty iterator if no matches.
+
+    Raises:
+      WildcardException: for bucket-only uri.
+    """
+    for bucket_listing_ref in self. __iter__():
+      if bucket_listing_ref.HasKey():
+        yield bucket_listing_ref.GetKey()
+
+  def IterUris(self):
+    """
+    Convenience iterator that runs underlying iterator and returns StorageUri
+    for each iteration.
+
+    Yields:
+      StorageUri, or empty iterator if no matches.
+    """
+    for bucket_listing_ref in self. __iter__():
+      yield bucket_listing_ref.GetUri()
+
+  def IterUrisForKeys(self):
+    """
+    Convenience iterator that runs underlying iterator and returns the
+    StorageUri for each iterated BucketListingRef that has a Key.
+
+    Yields:
+      StorageUri, or empty iterator if no matches.
+    """
+    for bucket_listing_ref in self. __iter__():
+      if bucket_listing_ref.HasKey():
+        yield bucket_listing_ref.GetUri()
+
+
+class FileWildcardIterator(WildcardIterator):
+  """WildcardIterator subclass for files and directories.
+
+  If you use recursive wildcards ('**') only a single such wildcard is
+  supported. For example you could use the wildcard '**/*.txt' to list all .txt
+  files in any subdirectory of the current directory, but you couldn't use a
+  wildcard like '**/abc/**/*.txt' (which would, if supported, let you find .txt
+  files in any subdirectory named 'abc').
+  """
+
+  def __init__(self, wildcard_uri, headers=None, debug=0):
+    """
+    Instantiate an iterator over BucketListingRefs matching given wildcard URI.
+
+    Args:
+      wildcard_uri: StorageUri that contains the wildcard to iterate.
+      headers: Dictionary containing optional HTTP headers to pass to boto.
+      debug: Debug level to pass in to boto connection (range 0..3).
+    """
+    self.wildcard_uri = wildcard_uri
+    self.headers = headers
+    self.debug = debug
+
+  def __iter__(self):
+    wildcard = self.wildcard_uri.object_name
+    match = re.search('\*\*', wildcard)
+    if match:
+      # Recursive wildcarding request ('.../**/...').
+      # Example input: wildcard = '/tmp/tmp2pQJAX/**/*'
+      base_dir = wildcard[:match.start()-1]
+      remaining_wildcard = wildcard[match.start()+2:]
+      # At this point for the above example base_dir = '/tmp/tmp2pQJAX' and
+      # remaining_wildcard = '/*'
+      if remaining_wildcard.startswith('*'):
+        raise WildcardException('Invalid wildcard with more than 2 consecutive '
+                                '*s (%s)' % wildcard)
+      # If there was no remaining wildcard past the recursive wildcard,
+      # treat it as if it were a '*'. For example, file://tmp/** is equivalent
+      # to file://tmp/**/*
+      if not remaining_wildcard:
+        remaining_wildcard = '*'
+      # Skip slash(es).
+      remaining_wildcard = remaining_wildcard.lstrip(os.sep)
+      filepaths = []
+      for dirpath, unused_dirnames, filenames in os.walk(base_dir):
+        filepaths.extend(
+            os.path.join(dirpath, f) for f in fnmatch.filter(filenames,
+                                                             remaining_wildcard)
+        )
+    else:
+      # Not a recursive wildcarding request.
+      filepaths = glob.glob(wildcard)
+    for filepath in filepaths:
+      expanded_uri = self.wildcard_uri.clone_replace_name(filepath)
+      yield BucketListingRef(expanded_uri)
+
+  def IterKeys(self):
+    """
+    Placeholder to allow polymorphic use of WildcardIterator.
+
+    Raises:
+      WildcardException: in all cases.
+    """
+    raise WildcardException(
+        'Iterating over Keys not possible for file wildcards')
+
+  def IterUris(self):
+    """
+    Convenience iterator that runs underlying iterator and returns StorageUri
+    for each iteration.
+
+    Yields:
+      StorageUri, or empty iterator if no matches.
+    """
+    for bucket_listing_ref in self. __iter__():
+      yield bucket_listing_ref.GetUri()
+
+
+class WildcardException(StandardError):
+  """Exception thrown for invalid wildcard URIs."""
+
+  def __init__(self, reason):
+    StandardError.__init__(self)
+    self.reason = reason
+
+  def __repr__(self):
+    return 'WildcardException: %s' % self.reason
+
+  def __str__(self):
+    return 'WildcardException: %s' % self.reason
+
+
+def wildcard_iterator(uri_or_str, proj_id_handler,
+                      bucket_storage_uri_class=BucketStorageUri,
+                      all_versions=False,
+                      headers=None, debug=0):
+  """Instantiate a WildCardIterator for the given StorageUri.
+
+  Args:
+    uri_or_str: StorageUri or URI string naming wildcard objects to iterate.
+    proj_id_handler: ProjectIdHandler to use for current command.
+    bucket_storage_uri_class: BucketStorageUri interface.
+        Settable for testing/mocking.
+    headers: Dictionary containing optional HTTP headers to pass to boto.
+    debug: Debug level to pass in to boto connection (range 0..3).
+
+  Returns:
+    A WildcardIterator that handles the requested iteration.
+  """
+
+  if isinstance(uri_or_str, basestring):
+    # Disable enforce_bucket_naming, to allow bucket names containing wildcard
+    # chars.
+    uri = boto.storage_uri(
+        uri_or_str, debug=debug, validate=False,
+        bucket_storage_uri_class=bucket_storage_uri_class,
+        suppress_consec_slashes=False)
+  else:
+    uri = uri_or_str
+
+  if uri.is_cloud_uri():
+    return CloudWildcardIterator(
+        uri, proj_id_handler,
+        bucket_storage_uri_class=bucket_storage_uri_class,
+        all_versions=all_versions,
+        headers=headers,
+        debug=debug)
+  elif uri.is_file_uri():
+    return FileWildcardIterator(uri, headers=headers, debug=debug)
+  else:
+    raise WildcardException('Unexpected type of StorageUri (%s)' % uri)
+
+
+def ContainsWildcard(uri_or_str):
+  """Checks whether uri_or_str contains a wildcard.
+
+  Args:
+    uri_or_str: StorageUri or URI string to check.
+
+  Returns:
+    bool indicator.
+  """
+  if isinstance(uri_or_str, basestring):
+    return bool(WILDCARD_REGEX.search(uri_or_str))
+  else:
+    return bool(WILDCARD_REGEX.search(uri_or_str.uri))