Syntax highlighting for C and Turtle code in documentation

1 files changed, 67 insertions, 0 deletions

diff --git a/lv2specgen/lv2specgen.py b/lv2specgen/lv2specgen.py
index 3c6db71..404545f 100755
--- a/lv2specgen/lv2specgen.py
+++ b/lv2specgen/lv2specgen.py
@@ -50,6 +50,8 @@ try:
     import pygments
     import pygments.lexers
     import pygments.formatters
+    from pygments.lexer import RegexLexer, include, bygroups
+    from pygments.token import Text, Comment, Operator, Keyword, Name, String, Literal
     have_pygments = True
 except ImportError:
     print("Error importing pygments, syntax highlighting disabled")
@@ -158,6 +160,58 @@ def getLabel(m, urinode):
     else:
         return ''
 
+if have_pygments:
+    # Based on sw.py by Philip Cooper
+    class Notation3Lexer(RegexLexer):
+        """
+        Lexer for N3 / Turtle / NT
+        """
+        name = 'N3'
+        aliases = ['n3', 'turtle']
+        filenames = ['*.n3', '*.ttl', '*.nt']
+        mimetypes = ['text/rdf+n3','application/x-turtle','application/n3']
+    
+        tokens = {
+            'comments': [
+                (r'(\s*#.*)', Comment)
+            ],
+            'root': [
+                include('comments'),
+                (r'(\s*@(?:prefix|base|keywords)\s*)(\w*:\s+)?(<[^> ]*>\s*\.\s*)',bygroups(Keyword,Name.Variable,Name.Namespace)),
+                (r'\s*(<[^>]*\>)', Name.Class, ('triple','predObj')),
+                (r'(\s*[a-zA-Z_:][a-zA-Z0-9\-_:]*\s)', Name.Class, ('triple','predObj')),
+                (r'\s*\[\]\s*', Name.Class, ('triple','predObj')),
+            ],
+            'triple' : [
+                (r'\s*\.\s*', Text, '#pop')
+            ],
+            'predObj': [
+                include('comments'),
+                (r'(\s*[a-zA-Z_:][a-zA-Z0-9\-_:]*\b\s*)', Operator, 'object'),
+                (r'\s*(<[^>]*\>)', Operator, 'object'),
+                (r'\s*\]\s*', Text, '#pop'),
+                (r'(?=\s*\.\s*)', Keyword, '#pop'), 
+            ],
+            'objList': [
+                include('comments'),
+                (r'\s*\)', Text, '#pop'),
+                include('object')
+            ],
+            'object': [
+                include('comments'),
+                (r'\s*\[', Text, 'predObj'),
+                (r'\s*<[^> ]*>', Name.Attribute),
+                (r'\s*("""(?:.|\n)*?""")(\@[a-z]{2-4}|\^\^<?[a-zA-Z0-9\-\:_#/\.]*>?)?\s*', bygroups(Literal.String,Text)),
+                (r'\s*".*?[^\\]"(?:\@[a-z]{2-4}|\^\^<?[a-zA-Z0-9\-\:_#/\.]*>?)?\s*', Literal.String),
+                (r'\s*[a-zA-Z0-9\-_\:]\s*', Name.Attribute),
+                (r'\s*\(', Text, 'objList'),
+                (r'\s*;\s*\n?', Text, '#pop'),
+                (r'\s*,\s*\n?', Text, '#pop'),  # Added by drobilla so "," is not an error
+                (r'(?=\s*\])', Text, '#pop'),            
+                (r'(?=\s*\.)', Text, '#pop'),           
+            ],
+        }
+    
 
 def getComment(m, urinode):
     c = findOne(m, urinode, lv2.documentation, None)
@@ -178,6 +232,19 @@ def getComment(m, urinode):
                     pygments.formatters.HtmlFormatter())
                 markup = code_rgx.sub(code_str, markup, 1)
 
+        # Syntax highlight all Turtle code
+        if have_pygments:
+            code_rgx = re.compile('<pre class="turtle-code">(.*?)</pre>', re.DOTALL)
+            while True:
+                code = code_rgx.search(markup)
+                if not code:
+                    break
+                match_str = xml.sax.saxutils.unescape(code.group(1))
+                code_str = pygments.highlight(
+                    match_str,
+                    Notation3Lexer(),
+                    pygments.formatters.HtmlFormatter())
+                markup = code_rgx.sub(code_str, markup, 1)
 
         # Add links to code documentation for identifiers
         rgx = re.compile('([^a-zA-Z0-9_:])(' + \