aboutsummaryrefslogtreecommitdiffstats
path: root/markdown_extensions/codehilite.py
diff options
context:
space:
mode:
authorYuri Takhteyev <yuri@freewisdom.org>2008-08-15 10:40:52 -0700
committerYuri Takhteyev <yuri@freewisdom.org>2008-08-15 10:40:52 -0700
commit5e3d88cf03303c52956f258898def2cd6a294674 (patch)
tree7899797058a8ee712664eaad062824b5c9e2f716 /markdown_extensions/codehilite.py
parent3a78823a59e9d0c89d541b2bd259ecfc55aef2cd (diff)
parent88c72d709be87321b736778d19d9277db02422a8 (diff)
downloadmarkdown-5e3d88cf03303c52956f258898def2cd6a294674.tar.gz
markdown-5e3d88cf03303c52956f258898def2cd6a294674.tar.bz2
markdown-5e3d88cf03303c52956f258898def2cd6a294674.zip
Merge branch 'master' of git@gitorious.org:python-markdown/mainline
Diffstat (limited to 'markdown_extensions/codehilite.py')
-rw-r--r--markdown_extensions/codehilite.py218
1 files changed, 218 insertions, 0 deletions
diff --git a/markdown_extensions/codehilite.py b/markdown_extensions/codehilite.py
new file mode 100644
index 0000000..73c1a79
--- /dev/null
+++ b/markdown_extensions/codehilite.py
@@ -0,0 +1,218 @@
+#!/usr/bin/python
+
+"""
+CodeHilite Extension for Python-Markdown
+=======================================
+
+Adds code/syntax highlighting to standard Python-Markdown code blocks.
+
+By [Waylan Limberg](http://achinghead.com/).
+
+Project website: http://achinghead.com/markdown-wikilinks/
+Contact: waylan [at] gmail [dot] com
+
+License: [BSD](http://www.opensource.org/licenses/bsd-license.php)
+
+Version: 0.2 (April 30, 2008)
+
+Dependencies:
+* [Python 2.3+](http://python.org/)
+* [Markdown 1.7+](http://www.freewisdom.org/projects/python-markdown/)
+* [Pygments](http://pygments.org/)
+
+"""
+
+import markdown
+
+# --------------- CONSTANTS YOU MIGHT WANT TO MODIFY -----------------
+
+try:
+ TAB_LENGTH = markdown.TAB_LENGTH
+except AttributeError:
+ TAB_LENGTH = 4
+
+
+# ------------------ The Main CodeHilite Class ----------------------
+class CodeHilite:
+ """
+ Determine language of source code, and pass it into the pygments hilighter.
+
+ Basic Usage:
+ >>> code = CodeHilite(src = text)
+ >>> html = code.hilite()
+
+ * src: Can be a string or any object with a .readline attribute.
+
+ * linenos: (Boolen) Turns line numbering 'on' or 'off' (off by default).
+
+ Low Level Usage:
+ >>> code = CodeHilite()
+ >>> code.src = text # String or anything with a .readline attribute
+ >>> code.linenos = True # True or False; Turns line numbering on or of.
+ >>> html = code.hilite()
+
+ """
+
+ def __init__(self, src=None, linenos = False):
+ self.src = src
+ self.lang = None
+ self.linenos = linenos
+
+ def hilite(self):
+ """
+ Pass code to the [Pygments](http://pygments.pocoo.org/) highliter with
+ optional line numbers. The output should then be styled with css to
+ your liking. No styles are applied by default - only styling hooks
+ (i.e.: <span class="k">).
+
+ returns : A string of html.
+
+ """
+
+ self.src = self.src.strip('\n')
+
+ self._getLang()
+
+ try:
+ from pygments import highlight
+ from pygments.lexers import get_lexer_by_name, guess_lexer, TextLexer
+ from pygments.formatters import HtmlFormatter
+ except ImportError:
+ # just escape and pass through
+ txt = self._escape(self.src)
+ '''if num:
+ txt = self._number(txt)
+ else :
+ txt = '<div class="codehilite"><pre>%s</pre></div>\n'% txt'''
+ txt = self._number(txt)
+ return txt
+ else:
+ try:
+ lexer = get_lexer_by_name(self.lang)
+ except ValueError:
+ try:
+ lexer = guess_lexer(self.src)
+ except ValueError:
+ lexer = TextLexer()
+ formatter = HtmlFormatter(linenos=self.linenos, cssclass="codehilite")
+ return highlight(self.src, lexer, formatter)
+
+ def _escape(self, txt):
+ """ basic html escaping """
+ txt = txt.replace('&', '&amp;')
+ txt = txt.replace('<', '&lt;')
+ txt = txt.replace('>', '&gt;')
+ txt = txt.replace('"', '&quot;')
+ return txt
+
+ def _number(self, txt):
+ """ Use <ol> for line numbering """
+ # Fix Whitespace
+ txt = txt.replace('\t', ' '*TAB_LENGTH)
+ txt = txt.replace(" "*4, "&nbsp; &nbsp; ")
+ txt = txt.replace(" "*3, "&nbsp; &nbsp;")
+ txt = txt.replace(" "*2, "&nbsp; ")
+
+ # Add line numbers
+ lines = txt.splitlines()
+ txt = '<div class="codehilite"><pre><ol>\n'
+ for line in lines:
+ txt += '\t<li>%s</li>\n'% line
+ txt += '</ol></pre></div>\n'
+ return txt
+
+
+ def _getLang(self):
+ """
+ Determines language of a code block from shebang lines and whether said
+ line should be removed or left in place. If the sheband line contains a
+ path (even a single /) then it is assumed to be a real shebang lines and
+ left alone. However, if no path is given (e.i.: #!python or :::python)
+ then it is assumed to be a mock shebang for language identifitation of a
+ code fragment and removed from the code block prior to processing for
+ code highlighting. When a mock shebang (e.i: #!python) is found, line
+ numbering is turned on. When colons are found in place of a shebang
+ (e.i.: :::python), line numbering is left in the current state - off
+ by default.
+
+ """
+
+ import re
+
+ #split text into lines
+ lines = self.src.split("\n")
+ #pull first line to examine
+ fl = lines.pop(0)
+
+ c = re.compile(r'''
+ (?:(?:::+)|(?P<shebang>[#]!)) #shebang or 2 or more colons
+ (?P<path>(?:/\w+)*[/ ])? # zero or 1 path ending in either a / or a single space
+ (?P<lang>\w*) # the language (a single / or space before lang is a path)
+ ''', re.VERBOSE)
+ # search first line for shebang
+ m = c.search(fl)
+ if m:
+ # we have a match
+ try:
+ self.lang = m.group('lang').lower()
+ except IndexError:
+ self.lang = None
+ if m.group('path'):
+ # path exists - restore first line
+ lines.insert(0, fl)
+ if m.group('shebang'):
+ # shebang exists - use line numbers
+ self.linenos = True
+ else:
+ # No match
+ lines.insert(0, fl)
+
+ self.src = "\n".join(lines).strip("\n")
+
+
+
+# ------------------ The Markdown Extension -------------------------------
+class CodeHiliteExtention(markdown.Extension):
+ def __init__(self, configs):
+ # define default configs
+ self.config = {
+ 'force_linenos' : [False, "Force line numbers - Default: False"]
+ }
+
+ # Override defaults with user settings
+ for key, value in configs:
+ # self.config[key][0] = value
+ self.setConfig(key, value)
+
+ def extendMarkdown(self, md, md_globals):
+
+ def _hiliteCodeBlock(parent_elem, lines, inList):
+ """
+ Overrides `_processCodeBlock` method in standard Markdown class
+ and sends code blocks to a code highlighting proccessor. The result
+ is then stored in the HtmlStash, a placeholder is inserted into
+ the dom and the remainder of the text file is processed recursively.
+
+ * parent_elem: DOM element to which the content will be added
+ * lines: a list of lines
+ * inList: a level
+
+ returns: None
+
+ """
+
+ detabbed, theRest = md.blockGuru.detectTabbed(lines)
+ text = "\n".join(detabbed).rstrip()+"\n"
+ code = CodeHilite(text, linenos=self.config['force_linenos'][0])
+ placeholder = md.htmlStash.store(code.hilite(), safe=True)
+ if parent_elem.text:
+ parent_elem.text += placeholder
+ else:
+ parent_elem.text = placeholder
+ md._processSection(parent_elem, theRest, inList)
+
+ md._processCodeBlock = _hiliteCodeBlock
+
+def makeExtension(configs={}):
+ return CodeHiliteExtention(configs=configs)
+