added mention of testing on windows

uses namedtemporaryfile

docs/source/detailed_installation.rst

@@ -24,11 +24,11 @@ configuration format is the same as used by Python’s configparser
 module, i.e. ::
 
     [library]
-    openjp2: /opt/openjp2-svn/lib/libopenjp2.so
+    openjp2: /somewhere/lib/libopenjp2.so
 
 This assumes, of course, that you've installed OpenJPEG into
-/opt/openjp2-svn on a linux system.  The location of the configuration file
-can vary as well (of course).  If you use either linux or mac, the path
+/opt/openjpeg on a linux system.  The location of the configuration file
+can vary as well.  If you use either linux or mac, the path
 to the configuration file would normally be ::
 
     $HOME/.config/glymur/glymurrc 
@@ -48,7 +48,7 @@ You may also include a line for the version 1.x openjpeg library if you have it
 installed in a non-standard place, i.e. ::
 
     [library]
-    openjpeg: /not/the/usual/location/lib/libopenjpeg.so
+    openjpeg: /somewhere/lib/libopenjpeg.so
 
 '''''''
 Testing

docs/source/how_do_i.rst

@@ -7,11 +7,11 @@ How do I...?
 ================
 Jp2k implements slicing via the :py:meth:`__getitem__` method, meaning that 
 multiple resolution imagery in a JPEG 2000 file can
-easily be accessed.  For example here's how to retrieve a full resolution and
-first lower-resolution image ::
+easily be accessed via array-style slicing.  For example here's how to
+retrieve a full resolution and first lower-resolution image ::
 
     >>> import glymur
-    >>> jp2file = glymur.data.nemo()
+    >>> jp2file = glymur.data.nemo() # just a path to a JPEG2000 file
     >>> jp2 = glymur.Jp2k(jp2file)
     >>> fullres = jp2[:]
     >>> print(fullres.shape)
@@ -20,8 +20,21 @@ first lower-resolution image ::
     >>> print(thumbnail.shape)
     (728, 1296, 3)
 
-The :py:meth:`read` method gives many more options for other JPEG 2000 features
-such as quality layers.
+The :py:meth:`read` method exposes many more options for other JPEG 2000
+features such as quality layers.
+
+... write images?
+=================
+So long as the image data can fit entirely into memory, array-style slicing may
+also be used to write JPEG 2000 files.
+    
+    >>> import glymur, numpy as np
+    >>> jp2 = glymur.Jp2k('zeros.jp2', mode='wb')
+    >>> jp2[:] = np.zeros((640, 480), dtype=np.uint8)
+
+The :py:meth:`write` method exposes many more options for other JPEG 2000
+features.  You should have OpenJPEG version 1.5 or more recent before writing
+JPEG 2000 images.
 
 ... display metadata?
 =====================
@@ -328,10 +341,10 @@ is currently limited to XML and UUID boxes.
 ... create an image with an alpha layer?
 ========================================
 
-OpenJPEG can create JP2 files with more than 3 components (requires
-the development version of OpenJPEG), but by default, any extra components are
-not described as such.  In order to do so, we need to rewrap such
-an image in a set of boxes that includes a channel definition box.
+OpenJPEG can create JP2 files with more than 3 components (use version 2.1.0+ 
+for this), but by default, any extra components are not described
+as such.  In order to do so, we need to rewrap such an image in a
+set of boxes that includes a channel definition box.
 
 This example is based on SciPy example code found at 
 http://scipy-lectures.github.io/advanced/image_processing/#basic-manipulations . 
@@ -389,7 +402,11 @@ Here's how the Preview application on the mac shows the RGBA image.
     
 ... work with XMP UUIDs?
 ========================
-XMP is metadata on steroids.  
+`Wikipedia <http://en.wikipedia.org/wiki/Extensible_Metadata_Platform>`_ states
+that "The Extensible Metadata Platform (XMP) is an ISO standard,
+originally created by Adobe Systems Inc., for the creation, processing
+and interchange of standardized and custom metadata for all kinds
+of resources."
 
 The example JP2 file shipped with glymur has an XMP UUID. ::
 

docs/source/introduction.rst

@@ -10,12 +10,11 @@ JPEG 2000 images is currently limited to images that can fit in memory.
 but it is strongly recommended to use version 2.1.0, which is the most recently 
 released version of OpenJPEG at this time.
 
-
 In regards to metadata, most JP2 boxes are properly interpreted.
 Certain optional JP2 boxes can also be written, including XML boxes and
 XMP UUIDs.  There is incomplete support for reading JPX metadata.
 
-Glymur 0.6 works on Python versions 2.7, 3.3 and 3.4.  If you have Python 2.6,
+Glymur works on Python versions 2.7, 3.3 and 3.4.  If you have Python 2.6,
 you should use the 0.5 series of Glymur.
 
 For more information about OpenJPEG, please consult http://www.openjpeg.org.

docs/source/roadmap.rst

@@ -2,8 +2,10 @@
 Platforms Tested (0.7.0 release only)
 -------------------------------------
     * Linux Mint 17 / Python 3.4.0 and 2.7.6 / OpenJPEG 2.1.0 and 1.3.0
-    * MacOS 10.6.8 / MacPorts / Python 3.4.1
+    * MacOS 10.6.8 / MacPorts Python 3.4.1, 3.3.5,and 2.7.8 / OpenJPEG 2.1.0
     * CentOS 6.5 / Anaconda Python 3.4.1 / OpenJPEG 1.3.0
+    * Fedora 20 i386 / Python 2.7.5 and 3.3.2 / OpenJPEG 1.5.1
+    * Windows 7 32bit / Anaconda Python 2.7.6 and 3.4.1 / OpenJPEG 2.1.0
 
 ------------
 Known Issues

glymur/jp2k.py

@@ -772,7 +772,7 @@ class Jp2k(Jp2kBox):
             # Should have a slice object where start = stop = step = None
             self.write(data)
         else:
-            msg = "Images currently must be written entirely at once."
+            msg = "Partial write operations are currently not allowed."
             raise TypeError(msg)
 
     def __getitem__(self, pargs):
@@ -832,9 +832,11 @@ class Jp2k(Jp2kBox):
         # one of them.
         step = rows_step
     
-        if np.log2(step) != np.floor(np.log2(step)):
+        # Check if the step size is a power of 2.
+        if np.abs(np.log2(step) - np.round(np.log2(step))) > 1e-6:
             msg = "Row and column strides must be powers of 2."
             raise IndexError(msg)
+        rlevel = np.int(np.round(np.log2(step)))
 
         if rows.start is None:
             rows_start = 0
@@ -857,7 +859,7 @@ class Jp2k(Jp2kBox):
             cols_stop = cols.stop
         
         area = (rows_start, cols_start, rows_stop, cols_stop)
-        data = self.read(area=area, rlevel=np.int(np.log2(step)))
+        data = self.read(area=area, rlevel=rlevel)
         if len(pargs) == 2:
             return data
 

glymur/test/test_config.py

@@ -71,6 +71,7 @@ class TestSuite(unittest.TestCase):
                     Jp2k(self.jp2file)
 
     @unittest.skipIf(WARNING_INFRASTRUCTURE_ISSUE, WARNING_INFRASTRUCTURE_MSG)
+    @unittest.skipIf(os.name == "nt", 'named temporary file issue on windows')
     def test_xdg_env_config_file_is_bad(self):
         """A non-existant library location should be rejected."""
         with tempfile.TemporaryDirectory() as tdir: