add is_gs_url(), split_gs_url() static methods to gs_utils.py

py/utils/gs_utils.py

 #!/usr/bin/python
 
 # pylint: disable=C0301
 """
 Copyright 2014 Google Inc.
 
 Use of this source code is governed by a BSD-style license that can be
 found in the LICENSE file.
 
 Utilities for accessing Google Cloud Storage, using the boto library (wrapper
@@ skipping 34 matching lines @@
 from boto.s3.connection import SubdomainCallingFormat
 from boto.s3.prefix import Prefix
 
 # How many files to upload at once, by default.
 # TODO(epoger): Is there a way to compute this intelligently?  To some extent
 # it is a function of how many cores are on the machine, and how many other
 # processes it is running; but it's probably more a function of how much time
 # each core sits idle waiting for network I/O to complete.
 DEFAULT_UPLOAD_THREADS = 10
 
+_GS_PREFIX = 'gs://'
+
 
 class AnonymousGSConnection(GSConnection):
   """GSConnection class that allows anonymous connections.
 
   The GSConnection class constructor in
   https://github.com/boto/boto/blob/develop/boto/gs/connection.py doesn't allow
   for anonymous connections (connections without credentials), so we have to
   override it.
   """
   def __init__(self):
@@ skipping 508 matching lines @@
     dirs = []
     files = []
     for item in items:
       t = type(item)
       if t is Key:
         files.append(item.name[prefix_length:])
       elif t is Prefix:
         dirs.append(item.name[prefix_length:-1])
     return (dirs, files)
 
+  @staticmethod
+  def is_gs_url(url):
+    """Returns True if url is a legal Google Storage URL ("gs://bucket/file").
+    """
+    try:
+      if url.lower().startswith(_GS_PREFIX) and len(url) > len(_GS_PREFIX):
+        return url[len(_GS_PREFIX)].isalnum()
+      else:
+        return False
+    except AttributeError:
+      return False
+
+  @staticmethod
+  def split_gs_url(url):
+    """Returns (bucket, filepath) corresponding to a legal Google Storage URL.
+
+    Raises AttributeError if the input URL is not a legal Google Storage URL.
+    """
+    if not GSUtils.is_gs_url(url):
+      raise AttributeError('"%s" is not a legal Google Storage URL' % url)
+    prefix_removed = url[len(_GS_PREFIX):]
+    pathsep_index = prefix_removed.find('/')
+    if pathsep_index < 0:
+      return (prefix_removed, '')
+    else:
+      return (prefix_removed[:pathsep_index],
+              prefix_removed[pathsep_index+1:].strip('/'))
+
   def _connect_to_bucket(self, bucket):
     """Returns a Bucket object we can use to access a particular bucket in GS.
 
     Params:
       bucket: name of the bucket (e.g., 'chromium-skia-gm'), or a Bucket
           object (in which case this param is just returned as-is)
     """
     if type(bucket) is Bucket:
       return bucket
     try:
@@ skipping 52 matching lines @@
 
 def _get_local_md5(path):
   """Returns the MD5 hash of a file on local disk."""
   hasher = hashlib.md5()
   with open(path, 'rb') as f:
     while True:
       data = f.read(64*1024)
       if not data:
         return hasher.hexdigest()
       hasher.update(data)