Merge "Replace utils.to_bytes() with strutils.to_bytes()"

4 files changed, 81 insertions, 77 deletions

diff --git a/nova/openstack/common/strutils.py b/nova/openstack/common/strutils.py
index cdf70cb20..62e547e60 100644
--- a/nova/openstack/common/strutils.py
+++ b/nova/openstack/common/strutils.py
@@ -19,18 +19,33 @@
 System-level utilities and helper functions.
 """
 
+import re
 import sys
+import unicodedata
 
 from nova.openstack.common.gettextutils import _
 
 
+# Used for looking up extensions of text
+# to their 'multiplied' byte amount
+BYTE_MULTIPLIERS = {
+    '': 1,
+    't': 1024 ** 4,
+    'g': 1024 ** 3,
+    'm': 1024 ** 2,
+    'k': 1024,
+}
+BYTE_REGEX = re.compile(r'(^-?\d+)(\D*)')
+
 TRUE_STRINGS = ('1', 't', 'true', 'on', 'y', 'yes')
 FALSE_STRINGS = ('0', 'f', 'false', 'off', 'n', 'no')
 
+SLUGIFY_STRIP_RE = re.compile(r"[^\w\s-]")
+SLUGIFY_HYPHENATE_RE = re.compile(r"[-\s]+")
+
 
 def int_from_bool_as_string(subject):
-    """
-    Interpret a string as a boolean and return either 1 or 0.
+    """Interpret a string as a boolean and return either 1 or 0.
 
     Any string value in:
 
@@ -44,8 +59,7 @@ def int_from_bool_as_string(subject):
 
 
 def bool_from_string(subject, strict=False):
-    """
-    Interpret a string as a boolean.
+    """Interpret a string as a boolean.
 
     A case-insensitive match is performed such that strings matching 't',
     'true', 'on', 'y', 'yes', or '1' are considered True and, when
@@ -78,9 +92,7 @@ def bool_from_string(subject, strict=False):
 
 
 def safe_decode(text, incoming=None, errors='strict'):
-    """
-    Decodes incoming str using `incoming` if they're
-    not already unicode.
+    """Decodes incoming str using `incoming` if they're not already unicode.
 
     :param incoming: Text's current encoding
     :param errors: Errors handling policy. See here for valid
@@ -119,11 +131,10 @@ def safe_decode(text, incoming=None, errors='strict'):
 
 def safe_encode(text, incoming=None,
                 encoding='utf-8', errors='strict'):
-    """
-    Encodes incoming str/unicode using `encoding`. If
-    incoming is not specified, text is expected to
-    be encoded with current python's default encoding.
-    (`sys.getdefaultencoding`)
+    """Encodes incoming str/unicode using `encoding`.
+
+    If incoming is not specified, text is expected to be encoded with
+    current python's default encoding. (`sys.getdefaultencoding`)
 
     :param incoming: Text's current encoding
     :param encoding: Expected encoding for text (Default UTF-8)
@@ -148,3 +159,58 @@ def safe_encode(text, incoming=None,
         return text.encode(encoding, errors)
 
     return text
+
+
+def to_bytes(text, default=0):
+    """Converts a string into an integer of bytes.
+
+    Looks at the last characters of the text to determine
+    what conversion is needed to turn the input text into a byte number.
+    Supports "B, K(B), M(B), G(B), and T(B)". (case insensitive)
+
+    :param text: String input for bytes size conversion.
+    :param default: Default return value when text is blank.
+
+    """
+    match = BYTE_REGEX.search(text)
+    if match:
+        magnitude = int(match.group(1))
+        mult_key_org = match.group(2)
+        if not mult_key_org:
+            return magnitude
+    elif text:
+        msg = _('Invalid string format: %s') % text
+        raise TypeError(msg)
+    else:
+        return default
+    mult_key = mult_key_org.lower().replace('b', '', 1)
+    multiplier = BYTE_MULTIPLIERS.get(mult_key)
+    if multiplier is None:
+        msg = _('Unknown byte multiplier: %s') % mult_key_org
+        raise TypeError(msg)
+    return magnitude * multiplier
+
+
+def to_slug(value, incoming=None, errors="strict"):
+    """Normalize string.
+
+    Convert to lowercase, remove non-word characters, and convert spaces
+    to hyphens.
+
+    Inspired by Django's `slugify` filter.
+
+    :param value: Text to slugify
+    :param incoming: Text's current encoding
+    :param errors: Errors handling policy. See here for valid
+        values http://docs.python.org/2/library/codecs.html
+    :returns: slugified unicode representation of `value`
+    :raises TypeError: If text is not an instance of basestring
+    """
+    value = safe_decode(value, incoming, errors)
+    # NOTE(aababilov): no need to use safe_(encode|decode) here:
+    # encodings are always "ascii", error handling is always "ignore"
+    # and types are always known (first: unicode; second: str)
+    value = unicodedata.normalize("NFKD", value).encode(
+        "ascii", "ignore").decode("ascii")
+    value = SLUGIFY_STRIP_RE.sub("", value).strip().lower()
+    return SLUGIFY_HYPHENATE_RE.sub("-", value)
diff --git a/nova/tests/test_utils.py b/nova/tests/test_utils.py
index f400bb899..609280261 100644
--- a/nova/tests/test_utils.py
+++ b/nova/tests/test_utils.py
@@ -38,41 +38,6 @@ from nova import utils
 CONF = cfg.CONF
 
 
-class ByteConversionTest(test.TestCase):
-    def test_string_conversions(self):
-        working_examples = {
-            '1024KB': 1048576,
-            '1024TB': 1125899906842624,
-            '1024K': 1048576,
-            '1024T': 1125899906842624,
-            '1TB': 1099511627776,
-            '1T': 1099511627776,
-            '1KB': 1024,
-            '1K': 1024,
-            '1B': 1,
-            '1B': 1,
-            '1': 1,
-            '1MB': 1048576,
-            '7MB': 7340032,
-            '0MB': 0,
-            '0KB': 0,
-            '0TB': 0,
-            '': 0,
-        }
-        for (in_value, expected_value) in working_examples.items():
-            b_value = utils.to_bytes(in_value)
-            self.assertEquals(expected_value, b_value)
-            if len(in_value):
-                in_value = "-" + in_value
-                b_value = utils.to_bytes(in_value)
-                self.assertEquals(expected_value * -1, b_value)
-        breaking_examples = [
-            'junk1KB', '1023BBBB',
-        ]
-        for v in breaking_examples:
-            self.assertRaises(TypeError, utils.to_bytes, v)
-
-
 class GetFromPathTestCase(test.TestCase):
     def test_tolerates_nones(self):
         f = utils.get_from_path
diff --git a/nova/utils.py b/nova/utils.py
index fe360cac6..5e968bd35 100644
--- a/nova/utils.py
+++ b/nova/utils.py
@@ -411,34 +411,6 @@ def utf8(value):
     return value
 
 
-def to_bytes(text, default=0):
-    """Try to turn a string into a number of bytes. Looks at the last
-    characters of the text to determine what conversion is needed to
-    turn the input text into a byte number.
-
-    Supports: B/b, K/k, M/m, G/g, T/t (or the same with b/B on the end)
-
-    """
-    # Take off everything not number 'like' (which should leave
-    # only the byte 'identifier' left)
-    mult_key_org = text.lstrip('-1234567890')
-    mult_key = mult_key_org.lower()
-    mult_key_len = len(mult_key)
-    if mult_key.endswith("b"):
-        mult_key = mult_key[0:-1]
-    try:
-        multiplier = BYTE_MULTIPLIERS[mult_key]
-        if mult_key_len:
-            # Empty cases shouldn't cause text[0:-0]
-            text = text[0:-mult_key_len]
-        return int(text) * multiplier
-    except KeyError:
-        msg = _('Unknown byte multiplier: %s') % mult_key_org
-        raise TypeError(msg)
-    except ValueError:
-        return default
-
-
 def get_from_path(items, path):
     """Returns a list of items matching the specified path.
 
diff --git a/nova/virt/images.py b/nova/virt/images.py
index b752b7a64..e1b42eafc 100755
--- a/nova/virt/images.py
+++ b/nova/virt/images.py
@@ -30,6 +30,7 @@ from nova import exception
 from nova.image import glance
 from nova.openstack.common import fileutils
 from nova.openstack.common import log as logging
+from nova.openstack.common import strutils
 from nova import utils
 
 LOG = logging.getLogger(__name__)
@@ -90,8 +91,8 @@ class QemuImgInfo(object):
         if real_size:
             details = real_size.group(1)
         try:
-            details = utils.to_bytes(details)
-        except (TypeError, ValueError):
+            details = strutils.to_bytes(details)
+        except TypeError:
             pass
         return details