From 2397ee84bf8ac27ed16942ba37dab317a334a68d Mon Sep 17 00:00:00 2001
From: "Marshall T. Vandegrift" <llasram@gmail.com>
Date: Sun, 11 Jan 2009 15:41:34 -0500
Subject: [PATCH] Mobipocket support:   - Add 'vertical-align: text-top' as a
 <sup/>-generating case   - Add titles to title-less known <guide/>
 <reference/>s and cleanly     ignoren any others   - Add transform to
 generate results of CSS 'text-transform' and     'font-variant'

---
 src/calibre/ebooks/mobi/mobiml.py             |   2 +-
 src/calibre/ebooks/mobi/writer.py             |   6 +-
 src/calibre/ebooks/oeb/base.py                |  23 +++-
 .../ebooks/oeb/transforms/manglecase.py       | 111 ++++++++++++++++++
 4 files changed, 139 insertions(+), 3 deletions(-)
 create mode 100644 src/calibre/ebooks/oeb/transforms/manglecase.py

diff --git a/src/calibre/ebooks/mobi/mobiml.py b/src/calibre/ebooks/mobi/mobiml.py
index 7dd2224bc3..73542da6c1 100644
--- a/src/calibre/ebooks/mobi/mobiml.py
+++ b/src/calibre/ebooks/mobi/mobiml.py
@@ -303,7 +303,7 @@ class MobiMLizer(object):
         else:
             istate.family = 'serif'
         valign = style['vertical-align']
-        if valign in ('super', 'sup') or asfloat(valign) > 0:
+        if valign in ('super', 'text-top') or asfloat(valign) > 0:
             istate.valign = 'super'
         elif valign == 'sub'  or asfloat(valign) < 0:
             istate.valign = 'sub'
diff --git a/src/calibre/ebooks/mobi/writer.py b/src/calibre/ebooks/mobi/writer.py
index baf4e11b2d..4e14c524cb 100644
--- a/src/calibre/ebooks/mobi/writer.py
+++ b/src/calibre/ebooks/mobi/writer.py
@@ -28,6 +28,7 @@ from calibre.ebooks.oeb.transforms.flatcss import CSSFlattener
 from calibre.ebooks.oeb.transforms.rasterize import SVGRasterizer
 from calibre.ebooks.oeb.transforms.trimmanifest import ManifestTrimmer
 from calibre.ebooks.oeb.transforms.htmltoc import HTMLTOCAdder
+from calibre.ebooks.oeb.transforms.manglecase import CaseMangler
 from calibre.ebooks.mobi.palmdoc import compress_doc
 from calibre.ebooks.mobi.langcodes import iana2mobi
 from calibre.ebooks.mobi.mobiml import MBP_NS, MBP, MobiMLizer
@@ -114,7 +115,8 @@ class Serializer(object):
         buffer.write('<guide>')
         for ref in self.oeb.guide.values():
             path, frag = urldefrag(ref.href)
-            if hrefs[path].media_type not in OEB_DOCS:
+            if hrefs[path].media_type not in OEB_DOCS or
+               not ref.title:
                 continue
             buffer.write('<reference title="%s" type="%s" '
                          % (ref.title, ref.type))
@@ -485,12 +487,14 @@ def main(argv=sys.argv):
     fbase = context.dest.fbase
     fkey = context.dest.fnums.values()
     tocadder = HTMLTOCAdder()
+    mangler = CaseMangler()
     flattener = CSSFlattener(
         fbase=fbase, fkey=fkey, unfloat=True, untable=True)
     rasterizer = SVGRasterizer()
     trimmer = ManifestTrimmer()
     mobimlizer = MobiMLizer()
     tocadder.transform(oeb, context)
+    mangler.transform(oeb, context)
     flattener.transform(oeb, context)
     rasterizer.transform(oeb, context)
     mobimlizer.transform(oeb, context)
diff --git a/src/calibre/ebooks/oeb/base.py b/src/calibre/ebooks/oeb/base.py
index 8cc8fc320e..178e8aba05 100644
--- a/src/calibre/ebooks/oeb/base.py
+++ b/src/calibre/ebooks/oeb/base.py
@@ -66,6 +66,25 @@ OEB_RASTER_IMAGES = set([GIF_MIME, JPEG_MIME, PNG_MIME])
 OEB_IMAGES = set([GIF_MIME, JPEG_MIME, PNG_MIME, SVG_MIME])
 
 MS_COVER_TYPE = 'other.ms-coverimage-standard'
+GUIDE_TITLES = {
+    'cover': 'Cover',
+    'title-page': 'Title Page',
+    'toc': 'Table of Contents',
+    'index': 'Index',
+    'glossary': 'Glossary',
+    'acknowledgements': 'Acknowledgements',
+    'bibliography': 'Bibliography',
+    'colophon': 'Colophon',
+    'copyright-page': 'Copyright',
+    'dedication': 'Dedication',
+    'epigraph': 'Epigraph',
+    'foreword': 'Foreword',
+    'loi': 'List of Illustrations',
+    'lot': 'List of Tables',
+    'notes': 'Notes',
+    'preface': 'Preface',
+    'text': 'Begin Reading'
+}
 
 recode = lambda s: s.decode('iso-8859-1').encode('ascii', 'xmlcharrefreplace')
 ENTITYDEFS = dict((k, recode(v)) for k, v in htmlentitydefs.entitydefs.items())
@@ -537,10 +556,12 @@ class Spine(object):
 class Guide(object):
     class Reference(object):
         def __init__(self, type, title, href):
+            if not title and type in GUIDE_TITLES:
+                title = GUIDE_TITLES[type]
             self.type = type
             self.title = title
             self.href = urlnormalize(href)
-
+        
         def __repr__(self):
             return 'Reference(type=%r, title=%r, href=%r)' \
                 % (self.type, self.title, self.href)
diff --git a/src/calibre/ebooks/oeb/transforms/manglecase.py b/src/calibre/ebooks/oeb/transforms/manglecase.py
new file mode 100644
index 0000000000..07fec10fb2
--- /dev/null
+++ b/src/calibre/ebooks/oeb/transforms/manglecase.py
@@ -0,0 +1,111 @@
+'''
+CSS case-mangling transform.
+'''
+from __future__ import with_statement
+
+__license__   = 'GPL v3'
+__copyright__ = '2008, Marshall T. Vandegrift <llasram@gmail.com>'
+
+import sys
+import os
+import re
+import operator
+import math
+from itertools import chain
+from collections import defaultdict
+from lxml import etree
+from calibre.ebooks.oeb.base import XHTML, XHTML_NS
+from calibre.ebooks.oeb.base import CSS_MIME
+from calibre.ebooks.oeb.base import namespace
+from calibre.ebooks.oeb.stylizer import Stylizer
+
+CASE_MANGLER_CSS = """
+.calibre_lowercase {
+    font-variant: normal;
+    font-size: 0.65em;
+}
+"""
+
+TEXT_TRANSFORMS = set(['capitalize', 'uppercase', 'lowercase'])
+
+class CaseMangler(object):
+    def transform(self, oeb, context):
+        oeb.logger.info('Applying case-transforming CSS...')
+        self.oeb = oeb
+        self.profile = context.source
+        self.mangle_spine()
+    
+    def mangle_spine(self):
+        id, href = self.oeb.manifest.generate('manglecase', 'manglecase.css')
+        self.oeb.manifest.add(id, href, CSS_MIME, data=CASE_MANGLER_CSS)
+        for item in self.oeb.spine:
+            html = item.data
+            etree.SubElement(html.find(XHTML('head')), XHTML('link'),
+                             rel='stylesheet', href=href, type=CSS_MIME)
+            stylizer = Stylizer(html, item.href, self.oeb, self.profile)
+            self.mangle_elem(html.find(XHTML('body')), stylizer)
+    
+    def text_transform(self, transform, text):
+        if transform == 'capitalize':
+            return text.title()
+        elif transform == 'uppercase':
+            return text.upper()
+        elif transform == 'lowercase':
+            return text.lower()
+        return text
+    
+    def split_text(self, text):
+        results = ['']
+        isupper = text[0].isupper()
+        for char in text:
+            if char.isupper() == isupper:
+                results[-1] += char
+            else:
+                isupper = not isupper
+                results.append(char)
+        return results
+    
+    def smallcaps_elem(self, elem, attr):
+        texts = self.split_text(getattr(elem, attr))
+        setattr(elem, attr, None)
+        last = elem if attr == 'tail' else None
+        attrib = {'class': 'calibre_lowercase'}
+        for text in texts:
+            if text.isupper():
+                if last is None:
+                    elem.text = text
+                else:
+                    last.tail = text
+            else:
+                child = etree.Element(XHTML('span'), attrib=attrib)
+                child.text = text.upper()
+                if last is None:
+                    elem.insert(0, child)
+                else:
+                    # addnext() moves the tail for some reason
+                    tail = last.tail
+                    last.addnext(child)
+                    last.tail = tail
+                    child.tail = None
+                last = child
+    
+    def mangle_elem(self, elem, stylizer):
+        if not isinstance(elem.tag, basestring) or \
+           namespace(elem.tag) != XHTML_NS:
+            return
+        children = list(elem)
+        style = stylizer.style(elem)
+        transform = style['text-transform']
+        variant = style['font-variant']
+        if elem.text:
+            if transform in TEXT_TRANSFORMS:
+                elem.text = self.text_transform(transform, elem.text)
+            if variant == 'small-caps':
+                self.smallcaps_elem(elem, 'text')
+        for child in children:
+            self.mangle_elem(child, stylizer)
+            if child.tail:
+                if transform in TEXT_TRANSFORMS:
+                    child.tail = self.text_transform(transform, child.tail)
+                if variant == 'small-caps':
+                    self.smallcaps_elem(child, 'tail')