Merge pull request #4655 from minrk/marked-nbconvert

prefer marked to pandoc for markdown2html

IPython/nbconvert/filters/marked.js

@@ -0,0 +1,54 b''
+// Node.js script for markdown to html conversion
+// This applies the same math extraction and marked settings
+// that we use in the live notebook.
+
+// IPython static_path dir relative to here:
+var static_path = __dirname + "/../../html/static/";
+
+var fs = require('fs');
+var IPython;
+// marked can be loaded with require,
+// the others must be execfiled
+var marked = require(static_path + 'components/marked/lib/marked.js');
+
+eval(fs.readFileSync(static_path + "components/highlight.js/build/highlight.pack.js", 'utf8'));
+eval(fs.readFileSync(static_path + "base/js/namespace.js", 'utf8'));
+
+eval(fs.readFileSync(static_path + "base/js/utils.js", 'utf8'));
+eval(fs.readFileSync(static_path + "notebook/js/mathjaxutils.js", 'utf8'));
+
+// this is copied from notebook.main. Should it be moved somewhere we can reuse it?
+marked.setOptions({
+    gfm : true,
+    tables: true,
+    langPrefix: "language-",
+    highlight: function(code, lang) {
+        if (!lang) {
+            // no language, no highlight
+            return code;
+        }
+        var highlighted;
+        try {
+            highlighted = hljs.highlight(lang, code, false);
+        } catch(err) {
+            highlighted = hljs.highlightAuto(code);
+        }
+        return highlighted.value;
+    }
+});
+
+// read the markdown from stdin
+var md='';
+process.stdin.on("data", function (data) {
+    md += data;
+});
+
+// perform the md2html transform once stdin is complete
+process.stdin.on("end", function () {
+    var text_and_math = IPython.mathjaxutils.remove_math(md);
+    var text = text_and_math[0];
+    var math = text_and_math[1];
+    var html = marked.parser(marked.lexer(text));
+    html = IPython.mathjaxutils.replace_math(html, math);
+    process.stdout.write(html);
+});

IPython/html/static/base/js/utils.js

@@ -448,6 +448,10 b' IPython.utils = (function (IPython) {'
 
     // http://stackoverflow.com/questions/2400935/browser-detection-in-javascript
     var browser = (function() {
+        if (typeof navigator === 'undefined') {
+            // navigator undefined in node
+            return 'None';
+        }
         var N= navigator.appName, ua= navigator.userAgent, tem;
         var M= ua.match(/(opera|chrome|safari|firefox|msie)\/?\s*(\.?\d+(\.\d+)*)/i);
         if (M && (tem= ua.match(/version\/([\.\d]+)/i))!= null) M[2]= tem[1];

IPython/html/static/notebook/js/mathjaxutils.js

@@ -106,12 +106,11 b' IPython.mathjaxutils = (function (IPython) {'
     //    math, then push the math string onto the storage array.
     //  The preProcess function is called on all blocks if it has been passed in
     var process_math = function (i, j, pre_process, math, blocks) {
-        var hub = MathJax.Hub;
         var block = blocks.slice(i, j + 1).join("").replace(/&/g, "&") // use HTML entity for &
         .replace(/</g, "&lt;") // use HTML entity for <
         .replace(/>/g, "&gt;") // use HTML entity for >
         ;
-        if (hub.Browser.isMSIE) {
+        if (IPython.utils.browser === 'msie') {
             block = block.replace(/(%[^\n]*)\n/g, "$1<br/>\n");
         }
         while (j > i) {
@@ -133,10 +132,6 b' IPython.mathjaxutils = (function (IPython) {'
     //    (which will be a paragraph).
     //
     var remove_math = function (text) {
-        if (!window.MathJax) {
-            return [text, null];
-        }
-
         var math = []; // stores math strings for later
         var start;
         var end;
@@ -241,9 +236,6 b' IPython.mathjaxutils = (function (IPython) {'
     //    and clear the math array (no need to keep it around).
     //
     var replace_math = function (text, math) {
-        if (!window.MathJax) {
-            return text;
-        }
         text = text.replace(/@@(\d+)@@/g, function (match, n) {
             return math[n];
         });

IPython/nbconvert/filters/markdown.py

@@ -16,21 +16,33 b' markdown within Jinja templates.'
 from __future__ import print_function
 
 # Stdlib imports
-import sys
+import os
 import subprocess
+from io import TextIOWrapper, BytesIO
 
+# IPython imports
 from IPython.nbconvert.utils.pandoc import pandoc
+from IPython.nbconvert.utils.exceptions import ConversionException
+from IPython.utils.process import find_cmd, FindCmdError
+from IPython.utils.py3compat import cast_bytes
 
 #-----------------------------------------------------------------------------
 # Functions
 #-----------------------------------------------------------------------------
+marked = os.path.join(os.path.dirname(__file__), "marked.js")
 
 __all__ = [
     'markdown2html',
+    'markdown2html_pandoc',
+    'markdown2html_marked',
     'markdown2latex',
-    'markdown2rst'
+    'markdown2rst',
 ]
 
+class NodeJSMissing(ConversionException):
+    """Exception raised when node.js is missing."""
+    pass
+
 def markdown2latex(source):
     """Convert a markdown string to LaTeX via pandoc.
 
@@ -49,11 +61,26 b' def markdown2latex(source):'
     """
     return pandoc(source, 'markdown', 'latex')
 
-
-def markdown2html(source):
+def markdown2html_pandoc(source):
     """Convert a markdown string to HTML via pandoc"""
     return pandoc(source, 'markdown', 'html', extra_args=['--mathjax'])
 
+def markdown2html_marked(source, encoding='utf-8'):
+    """Convert a markdown string to HTML via marked"""
+    command = ['node', marked]
+    try:
+        p = subprocess.Popen(command,
+                             stdin=subprocess.PIPE, stdout=subprocess.PIPE
+        )
+    except OSError as e:
+        raise NodeJSMissing(
+            "The command '%s' returned an error: %s.\n" % (" ".join(command), e) +
+            "Please check that Node.js is installed."
+        )
+    out, _ = p.communicate(cast_bytes(source, encoding))
+    out = TextIOWrapper(BytesIO(out), encoding, 'replace').read()
+    return out.rstrip('\n')
+
 def markdown2rst(source):
     """Convert a markdown string to LaTeX via pandoc.
 
@@ -72,3 +99,9 b' def markdown2rst(source):'
     """
     return pandoc(source, 'markdown', 'rst')
 
+try:
+    find_cmd('node')
+except FindCmdError:
+    markdown2html = markdown2html_pandoc
+else:
+    markdown2html = markdown2html_marked

setupbase.py

@@ -164,6 +164,7 b' def find_package_data():'
         'IPython.qt.console' : ['resources/icon/*.svg'],
         'IPython.nbconvert' : nbconvert_templates +
             ['tests/files/*.*', 'exporters/tests/files/*.*'],
+        'IPython.nbconvert.filters' : ['marked.js'],
         'IPython.nbformat' : ['tests/*.ipynb']
     }
     return package_data