From 71c59a6b9569fc2aafc4d4e41c21cca04b29cd65 Mon Sep 17 00:00:00 2001
From: wiemann <wiemann@929543f6-e4f2-0310-98a6-ba3bd3dd1d04>
Date: Mon, 20 Jun 2005 19:20:00 +0000
Subject: moved create_unimap.py to tools/dev/

git-svn-id: http://svn.code.sf.net/p/docutils/code/trunk/docutils@3534 929543f6-e4f2-0310-98a6-ba3bd3dd1d04
---
 tools/dev/create_unimap.py | 80 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 80 insertions(+)
 create mode 100755 tools/dev/create_unimap.py

(limited to 'tools/dev/create_unimap.py')

diff --git a/tools/dev/create_unimap.py b/tools/dev/create_unimap.py
new file mode 100755
index 000000000..260913ed3
--- /dev/null
+++ b/tools/dev/create_unimap.py
@@ -0,0 +1,80 @@
+#!/usr/bin/env python
+
+# Author: Felix Wiemann
+# Contact: Felix_Wiemann@ososo.de
+# Revision: $Revision$
+# Date: $Date$
+# Copyright: This file has been placed in the public domain.
+
+# Call: create_unimap.py < unicode.xml > unicode_latex.py
+#
+# Get unicode.xml from
+# <http://www.w3.org/2003/entities/xml/unicode.xml>.
+
+from xml.dom import minidom
+import sys
+import pprint
+
+def w(s):
+    if isinstance(s, unicode):
+        s = s.encode('utf8')
+    sys.stdout.write(s)
+
+text_map = {}
+math_map = {}
+
+class Visitor:
+    
+    """Node visitor for contents of unicode.xml."""
+
+    def visit_character(self, node):
+        for n in node.childNodes:
+            if n.nodeName == 'latex':
+                code = node.attributes['dec'].value
+                if '-' in code:
+                    # I don't know what this means, but we probably
+                    # don't need it....
+                    continue
+                if int(code) < 128:
+                    # Wrong (maps "-" to "$-$", which is too wide) and
+                    # unnecessary (maps "a" to "{a}").
+                    continue
+                latex_code = n.childNodes[0].nodeValue.encode('ascii').strip()
+                if node.attributes['mode'].value == 'math':
+                    math_map[unichr(int(code))] = '$%s$' % latex_code
+                else:
+                    text_map[unichr(int(code))] = '{%s}' % latex_code
+
+def call_visitor(node, visitor=Visitor()):
+    if isinstance(node, minidom.Text):
+        name = 'Text'
+    else:
+        name = node.nodeName.replace('#', '_')
+    if hasattr(visitor, 'visit_' + name):
+        getattr(visitor, 'visit_' + name)(node)
+    for child in node.childNodes:
+        call_visitor(child)
+    if hasattr(visitor, 'depart_' + name):
+        getattr(visitor, 'depart_' + name)(node)
+
+document = minidom.parse(sys.stdin)
+call_visitor(document)
+
+unicode_map = math_map
+unicode_map.update(text_map)
+# Now unicode_map contains the text entries plus dollar-enclosed math
+# entries for those chars for which no text entry exists.
+
+print '# Author: Felix Wiemann'
+print '# Contact: Felix_Wiemann@ososo.de'
+print '# Revision: $%s$' % 'Revision'
+print '# Date: $%s$' % 'Date'
+print '# Copyright: This file has been placed in the public domain.'
+print '#'
+print '# This is a mapping of Unicode characters to LaTeX'
+print '# equivalents.  The information has been extracted from'
+print '# <http://www.w3.org/2003/entities/xml/unicode.xml>.'
+print '# The extraction has been done by the "create_unimap.py"'
+print '# script written by Felix Wiemann.'
+print
+print 'unicode_map = %s' % pprint.pformat(unicode_map, indent=0)
-- 
cgit v1.2.1


From d3a60d28fa5ef090e66b3db54069811efc05afe6 Mon Sep 17 00:00:00 2001
From: wiemann <wiemann@929543f6-e4f2-0310-98a6-ba3bd3dd1d04>
Date: Mon, 20 Jun 2005 19:29:02 +0000
Subject: changed comment at the top of unicode_latex.py

git-svn-id: http://svn.code.sf.net/p/docutils/code/trunk/docutils@3535 929543f6-e4f2-0310-98a6-ba3bd3dd1d04
---
 tools/dev/create_unimap.py | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

(limited to 'tools/dev/create_unimap.py')

diff --git a/tools/dev/create_unimap.py b/tools/dev/create_unimap.py
index 260913ed3..1d1a2f8a0 100755
--- a/tools/dev/create_unimap.py
+++ b/tools/dev/create_unimap.py
@@ -70,11 +70,13 @@ print '# Contact: Felix_Wiemann@ososo.de'
 print '# Revision: $%s$' % 'Revision'
 print '# Date: $%s$' % 'Date'
 print '# Copyright: This file has been placed in the public domain.'
+print
+print '# This is a mapping of Unicode characters to LaTeX equivalents.'
+print '# The information has been extracted from'
+print '# <http://www.w3.org/2003/entities/xml/unicode.xml>, written by'
+print '# David Carlisle and Sebastian Rahtz.'
 print '#'
-print '# This is a mapping of Unicode characters to LaTeX'
-print '# equivalents.  The information has been extracted from'
-print '# <http://www.w3.org/2003/entities/xml/unicode.xml>.'
-print '# The extraction has been done by the "create_unimap.py"'
-print '# script written by Felix Wiemann.'
+print '# The extraction has been done by the "create_unimap.py" script'
+print '# located at <http://docutils.sf.net/tools/dev/create_unimap.py>.'
 print
 print 'unicode_map = %s' % pprint.pformat(unicode_map, indent=0)
-- 
cgit v1.2.1