Merge pull request #4696 from minrk/etree-fail

don't do anything if add_anchor fails

invalid content shouldn't necessarily kill nbconvert

IPython/nbconvert/filters/strings.py

@@ -1,207 +1,215 b''
 # coding: utf-8
 """String filters.
 
 Contains a collection of useful string manipulation filters for use in Jinja
 templates.
 """
 #-----------------------------------------------------------------------------
 # Copyright (c) 2013, the IPython Development Team.
 #
 # Distributed under the terms of the Modified BSD License.
 #
 # The full license is in the file COPYING.txt, distributed with this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import os
 import re
 import textwrap
 try:
     from urllib.parse import quote  # Py 3
 except ImportError:
     from urllib2 import quote  # Py 2
 from xml.etree import ElementTree
 
 from IPython.core.interactiveshell import InteractiveShell
 from IPython.utils import py3compat
 
 #-----------------------------------------------------------------------------
 # Functions
 #-----------------------------------------------------------------------------
 
 __all__ = [
     'wrap_text',
     'html2text',
     'add_anchor',
     'strip_dollars',
     'strip_files_prefix',
     'comment_lines',
     'get_lines',
     'ipython2python',
     'posix_path',
     'path2url',
     'add_prompts'
 ]
 
 
 def wrap_text(text, width=100):
     """ 
     Intelligently wrap text.
     Wrap text without breaking words if possible.
     
     Parameters
     ----------
     text : str
         Text to wrap.
     width : int, optional
         Number of characters to wrap to, default 100.
     """
 
     split_text = text.split('\n')
     wrp = map(lambda x:textwrap.wrap(x,width), split_text)
     wrpd = map('\n'.join, wrp)
     return '\n'.join(wrpd)
 
 
 def html2text(element):
     """extract inner text from html
     
     Analog of jQuery's $(element).text()
     """
     if isinstance(element, py3compat.string_types):
-        element = ElementTree.fromstring(element)
+        try:
+            element = ElementTree.fromstring(element)
+        except Exception:
+            # failed to parse, just return it unmodified
+            return element
     
     text = element.text or ""
     for child in element:
         text += html2text(child)
     text += (element.tail or "")
     return text
 
 
 def add_anchor(html):
     """Add an anchor-link to an html header tag
     
     For use in heading cells
     """
-    h = ElementTree.fromstring(py3compat.cast_bytes_py2(html, encoding='utf-8'))
+    try:
+        h = ElementTree.fromstring(py3compat.cast_bytes_py2(html, encoding='utf-8'))
+    except Exception:
+        # failed to parse, just return it unmodified
+        return html
     link = html2text(h).replace(' ', '-')
     h.set('id', link)
     a = ElementTree.Element("a", {"class" : "anchor-link", "href" : "#" + link})
     a.text = u'¶'
     h.append(a)
 
     # Known issue of Python3.x, ElementTree.tostring() returns a byte string
     # instead of a text string.  See issue http://bugs.python.org/issue10942
     # Workaround is to make sure the bytes are casted to a string.
     return py3compat.decode(ElementTree.tostring(h), 'utf-8')
 
 
 def add_prompts(code, first='>>> ', cont='... '):
     """Add prompts to code snippets"""
     new_code = []
     code_list = code.split('\n')
     new_code.append(first + code_list[0])
     for line in code_list[1:]:
         new_code.append(cont + line)
     return '\n'.join(new_code)
 
     
 def strip_dollars(text):
     """
     Remove all dollar symbols from text
     
     Parameters
     ----------
     text : str
         Text to remove dollars from
     """
 
     return text.strip('$')
 
 
 files_url_pattern = re.compile(r'(src|href)\=([\'"]?)/?files/')
 markdown_url_pattern = re.compile(r'(!?)\[(?P<caption>.*?)\]\(/?files/(?P<location>.*?)\)')
 
 def strip_files_prefix(text):
     """
     Fix all fake URLs that start with `files/`, stripping out the `files/` prefix.
     Applies to both urls (for html) and relative paths (for markdown paths).
     
     Parameters
     ----------
     text : str
         Text in which to replace 'src="files/real...' with 'src="real...'
     """
     cleaned_text = files_url_pattern.sub(r"\1=\2", text)
     cleaned_text = markdown_url_pattern.sub(r'\1[\2](\3)', cleaned_text)
     return cleaned_text
 
 
 def comment_lines(text, prefix='# '):
     """
     Build a Python comment line from input text.
     
     Parameters
     ----------
     text : str
         Text to comment out.
     prefix : str
         Character to append to the start of each line.
     """
     
     #Replace line breaks with line breaks and comment symbols.
     #Also add a comment symbol at the beginning to comment out
     #the first line.
     return prefix + ('\n'+prefix).join(text.split('\n')) 
 
 
 def get_lines(text, start=None,end=None):
     """
     Split the input text into separate lines and then return the 
     lines that the caller is interested in.
     
     Parameters
     ----------
     text : str
         Text to parse lines from.
     start : int, optional
         First line to grab from.
     end : int, optional
         Last line to grab from.
     """
     
     # Split the input into lines.
     lines = text.split("\n")
     
     # Return the right lines.
     return "\n".join(lines[start:end]) #re-join
 
 def ipython2python(code):
     """Transform IPython syntax to pure Python syntax
 
     Parameters
     ----------
 
     code : str
         IPython code, to be transformed to pure Python
     """
     shell = InteractiveShell.instance()
     return shell.input_transformer_manager.transform_cell(code)
 
 def posix_path(path):
     """Turn a path into posix-style path/to/etc
     
     Mainly for use in latex on Windows,
     where native Windows paths are not allowed.
     """
     if os.path.sep != '/':
         return path.replace(os.path.sep, '/')
     return path
 
 def path2url(path):
     """Turn a file path into a URL"""
     parts = path.split(os.path.sep)
     return '/'.join(quote(part) for part in parts)

IPython/nbconvert/filters/tests/test_strings.py

@@ -1,148 +1,153 b''
 """
 Module with tests for Strings
 """
 
 #-----------------------------------------------------------------------------
 # Copyright (c) 2013, the IPython Development Team.
 #
 # Distributed under the terms of the Modified BSD License.
 #
 # The full license is in the file COPYING.txt, distributed with this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 import os
 
 from ...tests.base import TestsBase
 from ..strings import (wrap_text, html2text, add_anchor, strip_dollars, 
     strip_files_prefix, get_lines, comment_lines, ipython2python, posix_path,
     add_prompts
 )
 
 
 #-----------------------------------------------------------------------------
 # Class
 #-----------------------------------------------------------------------------
 
 class TestStrings(TestsBase):
 
     def test_wrap_text(self):
         """wrap_text test"""
         test_text = """
         Tush! never tell me; I take it much unkindly
         That thou, Iago, who hast had my purse
         As if the strings were thine, shouldst know of this.
         """
         for length in [30,5,1]:
             self._confirm_wrap_text(test_text, length)
     
 
     def _confirm_wrap_text(self, text, length):
         for line in wrap_text(text, length).split('\n'):
             assert len(line) <= length
 
 
     def test_html2text(self):
         """html2text test"""
         #TODO: More tests
         self.assertEqual(html2text('<name>joe</name>'), 'joe')
         
 
     def test_add_anchor(self):
         """add_anchor test"""
         #TODO: More tests
         results = add_anchor('<b>Hello World!</b>')
         assert 'Hello World!' in results
         assert 'id="' in results
         assert 'class="anchor-link"' in results
         assert '<b' in results
         assert '</b>' in results
 
-        
+    def test_add_anchor_fail(self):
+        """add_anchor does nothing when it fails"""
+        html = '<h1>Hello <br>World!</h1>'
+        results = add_anchor(html)
+        self.assertEqual(html, results)
+
     def test_strip_dollars(self):
         """strip_dollars test"""
         tests = [
             ('', ''), 
             ('$$', ''), 
             ('$H$', 'H'), 
             ('$He', 'He'), 
             ('H$el', 'H$el'), 
             ('Hell$', 'Hell'),
             ('Hello', 'Hello'),
             ('W$o$rld', 'W$o$rld')]
         for test in tests:
             self._try_strip_dollars(test[0], test[1])
 
 
     def _try_strip_dollars(self, test, result):
         self.assertEqual(strip_dollars(test), result)
 
 
     def test_strip_files_prefix(self):
         """strip_files_prefix test"""
         tests = [
             ('', ''), 
             ('/files', '/files'), 
             ('test="/files"', 'test="/files"'), 
             ('My files are in `files/`', 'My files are in `files/`'),
             ('<a href="files/test.html">files/test.html</a>', '<a href="test.html">files/test.html</a>'),
             ('<a href="/files/test.html">files/test.html</a>', '<a href="test.html">files/test.html</a>'),
             ("<a href='files/test.html'>files/test.html</a>", "<a href='test.html'>files/test.html</a>"),
             ('<img src="files/url/location.gif">', '<img src="url/location.gif">'),
             ('<img src="/files/url/location.gif">', '<img src="url/location.gif">'),
             ('hello![caption]', 'hello![caption]'),
             ('hello![caption](/url/location.gif)', 'hello![caption](/url/location.gif)'),
             ('hello![caption](url/location.gif)', 'hello![caption](url/location.gif)'),
             ('hello![caption](url/location.gif)', 'hello![caption](url/location.gif)'),
             ('hello![caption](files/url/location.gif)', 'hello![caption](url/location.gif)'),
             ('hello![caption](/files/url/location.gif)', 'hello![caption](url/location.gif)'),
             ('hello [text](/files/url/location.gif)', 'hello [text](url/location.gif)'),
             ('hello [text space](files/url/location.gif)', 'hello [text space](url/location.gif)'),
         ]
         for test in tests:
             self._try_files_prefix(test[0], test[1])
 
 
     def _try_files_prefix(self, test, result):
         self.assertEqual(strip_files_prefix(test), result)
         
 
     def test_comment_lines(self):
         """comment_lines test"""
         for line in comment_lines('hello\nworld\n!').split('\n'):
             assert line.startswith('# ')
         for line in comment_lines('hello\nworld\n!', 'beep').split('\n'):
             assert line.startswith('beep')
         
 
     def test_get_lines(self):
         """get_lines test"""
         text = "hello\nworld\n!"
         self.assertEqual(get_lines(text, start=1), "world\n!")
         self.assertEqual(get_lines(text, end=2), "hello\nworld")
         self.assertEqual(get_lines(text, start=2, end=5), "!")
         self.assertEqual(get_lines(text, start=-2), "world\n!")
         
 
     def test_ipython2python(self):
         """ipython2python test"""
         #TODO: More tests
         results = ipython2python(u'%%pylab\nprint("Hello-World")').replace("u'", "'")
         self.fuzzy_compare(results, u"get_ipython().run_cell_magic('pylab', '', 'print(\"Hello-World\")')", 
             ignore_spaces=True, ignore_newlines=True)
     
     def test_posix_path(self):
         """posix_path test"""
         path_list = ['foo', 'bar']
         expected = '/'.join(path_list)
         native = os.path.join(*path_list)
         filtered = posix_path(native)
         self.assertEqual(filtered, expected)
     
     def test_add_prompts(self):
         """add_prompts test"""
         text1 = """for i in range(10):\n  i += 1\n  print i"""
         text2 = """>>> for i in range(10):\n...   i += 1\n...   print i"""
         self.assertEqual(text2, add_prompts(text1))