Merge pull request #3082 from minrk/image_garbage

A few small patches to image handling

IPython/core/display.py

@@ -1,557 +1,586 b''
 # -*- coding: utf-8 -*-
 """Top-level display functions for displaying object in different formats.
 
 Authors:
 
 * Brian Granger
 """
 
 #-----------------------------------------------------------------------------
-#       Copyright (C) 2008-2011 The IPython Development Team
+#       Copyright (C) 2013 The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 from __future__ import print_function
 
+import os
+
 from .displaypub import (
     publish_pretty, publish_html,
     publish_latex, publish_svg,
     publish_png, publish_json,
     publish_javascript, publish_jpeg
 )
 
 from IPython.utils.py3compat import string_types
 
 #-----------------------------------------------------------------------------
+# utility functions
+#-----------------------------------------------------------------------------
+
+def _safe_exists(path):
+    """check path, but don't let exceptions raise"""
+    try:
+        return os.path.exists(path)
+    except Exception:
+        return False
+
+#-----------------------------------------------------------------------------
 # Main functions
 #-----------------------------------------------------------------------------
 
 def display(*objs, **kwargs):
     """Display a Python object in all frontends.
 
     By default all representations will be computed and sent to the frontends.
     Frontends can decide which representation is used and how.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display.
     include : list or tuple, optional
         A list of format type strings (MIME types) to include in the
         format data dict. If this is set *only* the format types included
         in this list will be computed.
     exclude : list or tuple, optional
         A list of format type string (MIME types) to exclue in the format
         data dict. If this is set all format types will be computed,
         except for those included in this argument.
     """
     include = kwargs.get('include')
     exclude = kwargs.get('exclude')
 
     from IPython.core.interactiveshell import InteractiveShell
     inst = InteractiveShell.instance()
     format = inst.display_formatter.format
     publish = inst.display_pub.publish
 
     for obj in objs:
         format_dict = format(obj, include=include, exclude=exclude)
         publish('IPython.core.display.display', format_dict)
 
 
 def display_pretty(*objs, **kwargs):
     """Display the pretty (default) representation of an object.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw text data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_pretty(obj)
     else:
         display(*objs, include=['text/plain'])
 
 
 def display_html(*objs, **kwargs):
     """Display the HTML representation of an object.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw HTML data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_html(obj)
     else:
         display(*objs, include=['text/plain','text/html'])
 
 
 def display_svg(*objs, **kwargs):
     """Display the SVG representation of an object.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw svg data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_svg(obj)
     else:
         display(*objs, include=['text/plain','image/svg+xml'])
 
 
 def display_png(*objs, **kwargs):
     """Display the PNG representation of an object.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw png data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_png(obj)
     else:
         display(*objs, include=['text/plain','image/png'])
 
 
 def display_jpeg(*objs, **kwargs):
     """Display the JPEG representation of an object.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw JPEG data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_jpeg(obj)
     else:
         display(*objs, include=['text/plain','image/jpeg'])
 
 
 def display_latex(*objs, **kwargs):
     """Display the LaTeX representation of an object.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw latex data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_latex(obj)
     else:
         display(*objs, include=['text/plain','text/latex'])
 
 
 def display_json(*objs, **kwargs):
     """Display the JSON representation of an object.
     
     Note that not many frontends support displaying JSON.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw json data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_json(obj)
     else:
         display(*objs, include=['text/plain','application/json'])
 
 
 def display_javascript(*objs, **kwargs):
     """Display the Javascript representation of an object.
 
     Parameters
     ----------
     objs : tuple of objects
         The Python objects to display, or if raw=True raw javascript data to
         display.
     raw : bool
         Are the data objects raw data or Python objects that need to be
         formatted before display? [default: False]
     """
     raw = kwargs.pop('raw',False)
     if raw:
         for obj in objs:
             publish_javascript(obj)
     else:
         display(*objs, include=['text/plain','application/javascript'])
 
 #-----------------------------------------------------------------------------
 # Smart classes
 #-----------------------------------------------------------------------------
 
 
 class DisplayObject(object):
     """An object that wraps data to be displayed."""
 
     _read_flags = 'r'
 
     def __init__(self, data=None, url=None, filename=None):
         """Create a display object given raw data.
 
         When this object is returned by an expression or passed to the
         display function, it will result in the data being displayed
         in the frontend. The MIME type of the data should match the
         subclasses used, so the Png subclass should be used for 'image/png'
         data. If the data is a URL, the data will first be downloaded
         and then displayed. If
 
         Parameters
         ----------
         data : unicode, str or bytes
-            The raw data or a URL to download the data from.
+            The raw data or a URL or file to load the data from
         url : unicode
             A URL to download the data from.
         filename : unicode
             Path to a local file to load the data from.
         """
-        if data is not None and isinstance(data, string_types) and data.startswith('http'):
-            self.url = data
-            self.filename = None
-            self.data = None
-        else:
-            self.data = data
-            self.url = url
-            self.filename = None if filename is None else unicode(filename)
+        if data is not None and isinstance(data, string_types):
+            if data.startswith('http') and url is None:
+                url = data
+                filename = None
+                data = None
+            elif _safe_exists(data) and filename is None:
+                url = None
+                filename = data
+                data = None
+        
+        self.data = data
+        self.url = url
+        self.filename = None if filename is None else unicode(filename)
+        
         self.reload()
 
     def reload(self):
         """Reload the raw data from file or URL."""
         if self.filename is not None:
             with open(self.filename, self._read_flags) as f:
                 self.data = f.read()
         elif self.url is not None:
             try:
                 import urllib2
                 response = urllib2.urlopen(self.url)
                 self.data = response.read()
                 # extract encoding from header, if there is one:
                 encoding = None
                 for sub in response.headers['content-type'].split(';'):
                     sub = sub.strip()
                     if sub.startswith('charset'):
                         encoding = sub.split('=')[-1].strip()
                         break
                 # decode data, if an encoding was specified
                 if encoding:
                     self.data = self.data.decode(encoding, 'replace')
             except:
                 self.data = None
 
 class Pretty(DisplayObject):
 
     def _repr_pretty_(self):
         return self.data
 
 
 class HTML(DisplayObject):
 
     def _repr_html_(self):
         return self.data
 
 
 class Math(DisplayObject):
 
     def _repr_latex_(self):
         s = self.data.strip('$')
         return "$$%s$$" % s
 
 
 class Latex(DisplayObject):
 
     def _repr_latex_(self):
         return self.data
 
 
 class SVG(DisplayObject):
     
     # wrap data in a property, which extracts the <svg> tag, discarding
     # document headers
     _data = None
     
     @property
     def data(self):
         return self._data
     
     @data.setter
     def data(self, svg):
         if svg is None:
             self._data = None
             return
         # parse into dom object
         from xml.dom import minidom
         x = minidom.parseString(svg)
         # get svg tag (should be 1)
         found_svg = x.getElementsByTagName('svg')
         if found_svg:
             svg = found_svg[0].toxml()
         else:
             # fallback on the input, trust the user
             # but this is probably an error.
             pass
         self._data = svg
     
     def _repr_svg_(self):
         return self.data
 
 
 class JSON(DisplayObject):
 
     def _repr_json_(self):
         return self.data
 
 css_t = """$("head").append($("<link/>").attr({
   rel:  "stylesheet",
   type: "text/css",
   href: "%s"
 }));
 """
 
 lib_t1 = """$.getScript("%s", function () {
 """
 lib_t2 = """});
 """
 
 class Javascript(DisplayObject):
 
     def __init__(self, data=None, url=None, filename=None, lib=None, css=None):
         """Create a Javascript display object given raw data.
 
         When this object is returned by an expression or passed to the
         display function, it will result in the data being displayed
         in the frontend. If the data is a URL, the data will first be
         downloaded and then displayed. 
         
         In the Notebook, the containing element will be available as `element`,
         and jQuery will be available.  The output area starts hidden, so if
         the js appends content to `element` that should be visible, then
         it must call `container.show()` to unhide the area.
 
         Parameters
         ----------
         data : unicode, str or bytes
             The Javascript source code or a URL to download it from.
         url : unicode
             A URL to download the data from.
         filename : unicode
             Path to a local file to load the data from.
         lib : list or str
             A sequence of Javascript library URLs to load asynchronously before
             running the source code. The full URLs of the libraries should
             be given. A single Javascript library URL can also be given as a
             string.
         css: : list or str
             A sequence of css files to load before running the source code.
             The full URLs of the css files should be give. A single css URL
             can also be given as a string.
         """
         if isinstance(lib, basestring):
             lib = [lib]
         elif lib is None:
             lib = []
         if isinstance(css, basestring):
             css = [css]
         elif css is None:
             css = []
         if not isinstance(lib, (list,tuple)):
             raise TypeError('expected sequence, got: %r' % lib)
         if not isinstance(css, (list,tuple)):
             raise TypeError('expected sequence, got: %r' % css)
         self.lib = lib
         self.css = css
         super(Javascript, self).__init__(data=data, url=url, filename=filename)
 
     def _repr_javascript_(self):
         r = ''
         for c in self.css:
             r += css_t % c
         for l in self.lib:
             r += lib_t1 % l
         r += self.data
         r += lib_t2*len(self.lib)
         return r
 
+# constants for identifying png/jpeg data
+_PNG = b'\x89PNG\r\n\x1a\n'
+_JPEG = b'\xff\xd8'
 
 class Image(DisplayObject):
 
     _read_flags = 'rb'
     _FMT_JPEG = u'jpeg'
     _FMT_PNG = u'png'
     _ACCEPTABLE_EMBEDDINGS = [_FMT_JPEG, _FMT_PNG]
 
     def __init__(self, data=None, url=None, filename=None, format=u'png', embed=None, width=None, height=None):
         """Create a display an PNG/JPEG image given raw data.
 
         When this object is returned by an expression or passed to the
         display function, it will result in the image being displayed
         in the frontend.
 
         Parameters
         ----------
         data : unicode, str or bytes
             The raw data or a URL to download the data from.
         url : unicode
             A URL to download the data from.
         filename : unicode
             Path to a local file to load the data from.
         format : unicode
             The format of the image data (png/jpeg/jpg). If a filename or URL is given
             for format will be inferred from the filename extension.
         embed : bool
             Should the image data be embedded using a data URI (True) or be
             loaded using an <img> tag. Set this to True if you want the image
             to be viewable later with no internet connection in the notebook.
 
             Default is `True`, unless the keyword argument `url` is set, then
             default value is `False`.
 
             Note that QtConsole is not able to display images if `embed` is set to `False`
         width : int
             Width to which to constrain the image in html
         height : int
             Height to which to constrain the image in html
 
         Examples
         --------
         # embed implicitly True, works in qtconsole and notebook
         Image('http://www.google.fr/images/srpr/logo3w.png')
 
         # embed implicitly False, does not works in qtconsole but works in notebook if
         # internet connection available
         Image(url='http://www.google.fr/images/srpr/logo3w.png')
 
         """
         if filename is not None:
             ext = self._find_ext(filename)
         elif url is not None:
             ext = self._find_ext(url)
         elif data is None:
             raise ValueError("No image data found. Expecting filename, url, or data.")
-        elif isinstance(data, string_types) and data.startswith('http'):
+        elif isinstance(data, string_types) and (
+            data.startswith('http') or _safe_exists(data)
+        ):
             ext = self._find_ext(data)
         else:
             ext = None
 
         if ext is not None:
             format = ext.lower()
             if ext == u'jpg' or ext == u'jpeg':
                 format = self._FMT_JPEG
             if ext == u'png':
                 format = self._FMT_PNG
+        elif isinstance(data, bytes) and format == 'png':
+            # infer image type from image data header,
+            # only if format might not have been specified.
+            if data[:2] == _JPEG:
+                format = 'jpeg'
 
         self.format = unicode(format).lower()
         self.embed = embed if embed is not None else (url is None)
 
         if self.embed and self.format not in self._ACCEPTABLE_EMBEDDINGS:
             raise ValueError("Cannot embed the '%s' image format" % (self.format))
         self.width = width
         self.height = height
         super(Image, self).__init__(data=data, url=url, filename=filename)
 
     def reload(self):
         """Reload the raw data from file or URL."""
         if self.embed:
             super(Image,self).reload()
 
     def _repr_html_(self):
         if not self.embed:
             width = height = ''
             if self.width:
                 width = ' width="%d"' % self.width
             if self.height:
                 height = ' height="%d"' % self.height
             return u'<img src="%s"%s%s/>' % (self.url, width, height)
 
     def _repr_png_(self):
         if self.embed and self.format == u'png':
             return self.data
 
     def _repr_jpeg_(self):
         if self.embed and (self.format == u'jpeg' or self.format == u'jpg'):
             return self.data
 
     def _find_ext(self, s):
         return unicode(s.split('.')[-1].lower())
 
 
 def clear_output(stdout=True, stderr=True, other=True):
     """Clear the output of the current cell receiving output.
     
     Optionally, each of stdout/stderr or other non-stream data (e.g. anything
     produced by display()) can be excluded from the clear event.
     
     By default, everything is cleared.
     
     Parameters
     ----------
     stdout : bool [default: True]
         Whether to clear stdout.
     stderr : bool [default: True]
         Whether to clear stderr.
     other : bool [default: True]
         Whether to clear everything else that is not stdout/stderr
         (e.g. figures,images,HTML, any result of display()).
     """
     from IPython.core.interactiveshell import InteractiveShell
     if InteractiveShell.initialized():
         InteractiveShell.instance().display_pub.clear_output(
             stdout=stdout, stderr=stderr, other=other,
         )
     else:
         from IPython.utils import io
         if stdout:
             print('\033[2K\r', file=io.stdout, end='')
             io.stdout.flush()
         if stderr:
             print('\033[2K\r', file=io.stderr, end='')
             io.stderr.flush()
         

IPython/utils/jsonutil.py

@@ -1,211 +1,224 b''
 """Utilities to manipulate JSON objects.
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING.txt, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 # stdlib
 import math
 import re
 import types
 from datetime import datetime
 
 try:
     # base64.encodestring is deprecated in Python 3.x
     from base64 import encodebytes
 except ImportError:
     # Python 2.x
     from base64 import encodestring as encodebytes
 
 from IPython.utils import py3compat
 from IPython.utils.encoding import DEFAULT_ENCODING
 next_attr_name = '__next__' if py3compat.PY3 else 'next'
 
 #-----------------------------------------------------------------------------
 # Globals and constants
 #-----------------------------------------------------------------------------
 
 # timestamp formats
 ISO8601="%Y-%m-%dT%H:%M:%S.%f"
 ISO8601_PAT=re.compile(r"^\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}\.\d+$")
 
 #-----------------------------------------------------------------------------
 # Classes and functions
 #-----------------------------------------------------------------------------
 
 def rekey(dikt):
     """Rekey a dict that has been forced to use str keys where there should be
     ints by json."""
     for k in dikt.iterkeys():
         if isinstance(k, basestring):
             ik=fk=None
             try:
                 ik = int(k)
             except ValueError:
                 try:
                     fk = float(k)
                 except ValueError:
                     continue
             if ik is not None:
                 nk = ik
             else:
                 nk = fk
             if nk in dikt:
                 raise KeyError("already have key %r"%nk)
             dikt[nk] = dikt.pop(k)
     return dikt
 
 
 def extract_dates(obj):
     """extract ISO8601 dates from unpacked JSON"""
     if isinstance(obj, dict):
         obj = dict(obj) # don't clobber
         for k,v in obj.iteritems():
             obj[k] = extract_dates(v)
     elif isinstance(obj, (list, tuple)):
         obj = [ extract_dates(o) for o in obj ]
     elif isinstance(obj, basestring):
         if ISO8601_PAT.match(obj):
             obj = datetime.strptime(obj, ISO8601)
     return obj
 
 def squash_dates(obj):
     """squash datetime objects into ISO8601 strings"""
     if isinstance(obj, dict):
         obj = dict(obj) # don't clobber
         for k,v in obj.iteritems():
             obj[k] = squash_dates(v)
     elif isinstance(obj, (list, tuple)):
         obj = [ squash_dates(o) for o in obj ]
     elif isinstance(obj, datetime):
         obj = obj.strftime(ISO8601)
     return obj
 
 def date_default(obj):
     """default function for packing datetime objects in JSON."""
     if isinstance(obj, datetime):
         return obj.strftime(ISO8601)
     else:
         raise TypeError("%r is not JSON serializable"%obj)
 
 
 # constants for identifying png/jpeg data
 PNG = b'\x89PNG\r\n\x1a\n'
+# front of PNG base64-encoded
+PNG64 = b'iVBORw0KG'
 JPEG = b'\xff\xd8'
+# front of JPEG base64-encoded
+JPEG64 = b'/9'
 
 def encode_images(format_dict):
     """b64-encodes images in a displaypub format dict
 
     Perhaps this should be handled in json_clean itself?
 
     Parameters
     ----------
 
     format_dict : dict
         A dictionary of display data keyed by mime-type
 
     Returns
     -------
 
     format_dict : dict
         A copy of the same dictionary,
         but binary image data ('image/png' or 'image/jpeg')
         is base64-encoded.
 
     """
     encoded = format_dict.copy()
+
     pngdata = format_dict.get('image/png')
-    if isinstance(pngdata, bytes) and pngdata[:8] == PNG:
-        encoded['image/png'] = encodebytes(pngdata).decode('ascii')
+    if isinstance(pngdata, bytes):
+        # make sure we don't double-encode
+        if not pngdata.startswith(PNG64):
+            pngdata = encodebytes(pngdata)
+        encoded['image/png'] = pngdata.decode('ascii')
+
     jpegdata = format_dict.get('image/jpeg')
-    if isinstance(jpegdata, bytes) and jpegdata[:2] == JPEG:
-        encoded['image/jpeg'] = encodebytes(jpegdata).decode('ascii')
+    if isinstance(jpegdata, bytes):
+        # make sure we don't double-encode
+        if not jpegdata.startswith(JPEG64):
+            jpegdata = encodebytes(jpegdata)
+        encoded['image/jpeg'] = jpegdata.decode('ascii')
+
     return encoded
 
 
 def json_clean(obj):
     """Clean an object to ensure it's safe to encode in JSON.
 
     Atomic, immutable objects are returned unmodified.  Sets and tuples are
     converted to lists, lists are copied and dicts are also copied.
 
     Note: dicts whose keys could cause collisions upon encoding (such as a dict
     with both the number 1 and the string '1' as keys) will cause a ValueError
     to be raised.
 
     Parameters
     ----------
     obj : any python object
 
     Returns
     -------
     out : object
 
       A version of the input which will not cause an encoding error when
       encoded as JSON.  Note that this function does not *encode* its inputs,
       it simply sanitizes it so that there will be no encoding errors later.
 
     Examples
     --------
     >>> json_clean(4)
     4
     >>> json_clean(range(10))
     [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
     >>> sorted(json_clean(dict(x=1, y=2)).items())
     [('x', 1), ('y', 2)]
     >>> sorted(json_clean(dict(x=1, y=2, z=[1,2,3])).items())
     [('x', 1), ('y', 2), ('z', [1, 2, 3])]
     >>> json_clean(True)
     True
     """
     # types that are 'atomic' and ok in json as-is.  bool doesn't need to be
     # listed explicitly because bools pass as int instances
     atomic_ok = (unicode, int, types.NoneType)
 
     # containers that we need to convert into lists
     container_to_list = (tuple, set, types.GeneratorType)
 
     if isinstance(obj, float):
         # cast out-of-range floats to their reprs
         if math.isnan(obj) or math.isinf(obj):
             return repr(obj)
         return obj
 
     if isinstance(obj, atomic_ok):
         return obj
 
     if isinstance(obj, bytes):
         return obj.decode(DEFAULT_ENCODING, 'replace')
 
     if isinstance(obj, container_to_list) or (
         hasattr(obj, '__iter__') and hasattr(obj, next_attr_name)):
         obj = list(obj)
 
     if isinstance(obj, list):
         return [json_clean(x) for x in obj]
 
     if isinstance(obj, dict):
         # First, validate that the dict won't lose data in conversion due to
         # key collisions after stringification.  This can happen with keys like
         # True and 'true' or 1 and '1', which collide in JSON.
         nkeys = len(obj)
         nkeys_collapsed = len(set(map(str, obj)))
         if nkeys != nkeys_collapsed:
             raise ValueError('dict can not be safely converted to JSON: '
                              'key collision would lead to dropped values')
         # If all OK, proceed by making the new dict that will be json-safe
         out = {}
         for k,v in obj.iteritems():
             out[str(k)] = json_clean(v)
         return out
 
     # If we get here, we don't know how to handle the object, so we just get
     # its repr and return that.  This will catch lambdas, open sockets, class
     # objects, and any other complicated contraption that json can't encode
     return repr(obj)