From 0bda2a73bd83d1e0b4358acdef4449e3edb41d59 Mon Sep 17 00:00:00 2001
From: Rato <rato.aq2@gmail.com>
Date: Thu, 20 Jul 2017 22:23:09 +0200
Subject: [PATCH] Fix for #481: Enhance  output value

---
 HISTORY.rst                      |  1 +
 docs/properties.rst              |  4 ++-
 guessit/__init__.py              |  1 +
 guessit/rules/common/quantity.py | 55 ++++++++++++++++++++++++++++++++
 guessit/rules/properties/size.py | 10 ++----
 5 files changed, 63 insertions(+), 8 deletions(-)
 create mode 100644 guessit/rules/common/quantity.py

diff --git a/HISTORY.rst b/HISTORY.rst
index ebbec41..e30aa9e 100644
--- a/HISTORY.rst
+++ b/HISTORY.rst
@@ -97,6 +97,7 @@ History
 - Added default and configurable list of allowed languages and countries
 - Added `VC-1` as new `video_codec` possible value
 - Enhanced dash-separated `release_group` detection.
+- Changed `size` output to return `guessit.Quantity` object.
 
 2.1.4 (2017-06-01)
 ------------------
diff --git a/docs/properties.rst b/docs/properties.rst
index a5c07ea..2c58dae 100644
--- a/docs/properties.rst
+++ b/docs/properties.rst
@@ -265,7 +265,9 @@ Other properties
 
 - **size**
 
-  Size (MB, GB, TB).
+  Size (MB, GB, TB). Examples: ``1.2GB`` (``<Quantity [1.2GB]>``), ``430MB`` (``<Quantity [430MB]>``).
+
+  - ``[<guessit.Quantity>]`` (object has ``magnitude`` and ``units``)
 
 
 - **edition**
diff --git a/guessit/__init__.py b/guessit/__init__.py
index 22e9dbb..4b519c9 100644
--- a/guessit/__init__.py
+++ b/guessit/__init__.py
@@ -5,5 +5,6 @@ Extracts as much information as possible from a video file.
 """
 from .api import guessit, GuessItApi
 from .options import ConfigurationException
+from .rules.common.quantity import Quantity
 
 from .__version__ import __version__
diff --git a/guessit/rules/common/quantity.py b/guessit/rules/common/quantity.py
new file mode 100644
index 0000000..57614af
--- /dev/null
+++ b/guessit/rules/common/quantity.py
@@ -0,0 +1,55 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+"""
+Quantity
+"""
+import re
+
+import six
+
+
+class Quantity(object):
+    """
+    Represent a quantity object with magnitude and units.
+    """
+
+    parser_re = re.compile(r'(?P<magnitude>\d+(?:[.]\d+)?)(?P<units>[^\d]+)')
+
+    def __init__(self, magnitude, units):
+        self.magnitude = magnitude
+        self.units = units
+
+    @classmethod
+    def fromstring(cls, string):
+        """
+        Parse the string into a quantity object.
+        :param string:
+        :return:
+        """
+        values = cls.parser_re.match(string).groupdict()
+        try:
+            magnitude = int(values['magnitude'])
+        except ValueError:
+            magnitude = float(values['magnitude'])
+        units = values['units'].upper()
+
+        return Quantity(magnitude, units)
+
+    def __hash__(self):
+        return hash(str(self))
+
+    def __eq__(self, other):
+        if isinstance(other, six.string_types):
+            return str(self) == other
+        if not isinstance(other, Quantity):
+            return NotImplemented
+        return self.magnitude == other.magnitude and self.units == other.units
+
+    def __ne__(self, other):
+        return not self == other
+
+    def __repr__(self):
+        return '<Quantity [{0}]>'.format(self)
+
+    def __str__(self):
+        return '{0}{1}'.format(self.magnitude, self.units)
diff --git a/guessit/rules/properties/size.py b/guessit/rules/properties/size.py
index 84f0303..2657dd2 100644
--- a/guessit/rules/properties/size.py
+++ b/guessit/rules/properties/size.py
@@ -7,8 +7,9 @@ import re
 
 from rebulk import Rebulk
 
-from ..common.validators import seps_surround
 from ..common import dash
+from ..common.quantity import Quantity
+from ..common.validators import seps_surround
 
 
 def size():
@@ -17,13 +18,8 @@ def size():
     :return: Created Rebulk object
     :rtype: Rebulk
     """
-
-    def format_size(value):
-        """Format size using uppercase and no space."""
-        return re.sub(r'(?<=\d)[.](?=[^\d])', '', value.upper())
-
     rebulk = Rebulk().regex_defaults(flags=re.IGNORECASE, abbreviations=[dash])
     rebulk.defaults(name='size', validator=seps_surround)
-    rebulk.regex(r'\d+\.?[mgt]b', r'\d+\.\d+[mgt]b', formatter=format_size, tags=['release-group-prefix'])
+    rebulk.regex(r'\d+\.?[mgt]b', r'\d+\.\d+[mgt]b', formatter=Quantity.fromstring, tags=['release-group-prefix'])
 
     return rebulk