Fixes #2631: losing whitespace in epub/mobi stanzas.

[librarian.git] / librarian / pdf.py

diff --git a/librarian/pdf.py b/librarian/pdf.py
index 02438a6..9fb92b1 100644 (file)
--- a/librarian/pdf.py
+++ b/librarian/pdf.py
@@ -3,6 +3,12 @@
 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
+"""PDF creation library.
+
+Creates one big XML from the book and its children, converts it to LaTeX
+with TeXML, then runs it by XeLaTeX.
+
+"""
 from __future__ import with_statement
 import os
 import os.path
@@ -21,6 +27,7 @@ from librarian.dcparser import Person
 from librarian.parser import WLDocument
 from librarian import ParseError, DCNS, get_resource, OutputFile
 from librarian import functions
+from librarian.cover import WLCover
 
 
 functions.reg_substitute_entities()
@@ -33,13 +40,14 @@ STYLESHEETS = {
     'wl2tex': 'pdf/wl2tex.xslt',
 }
 
-CUSTOMIZATIONS = [
-    'nofootnotes',
-    'nothemes',
-    'onehalfleading',
-    'doubleleading',
-    'nowlfont',
-    ]
+#CUSTOMIZATIONS = [
+#    'nofootnotes',
+#    'nothemes',
+#    'defaultleading',
+#    'onehalfleading',
+#    'doubleleading',
+#    'nowlfont',
+#    ]
 
 def insert_tags(doc, split_re, tagname, exclude=None):
     """ inserts <tagname> for every occurence of `split_re' in text nodes in the `doc' tree
@@ -133,9 +141,13 @@ def hack_motifs(doc):
 
 
 def parse_creator(doc):
-    """ find all dc:creator and dc.contributor tags and add *_parsed versions with forenames first """
+    """Generates readable versions of creator and translator tags.
+
+    Finds all dc:creator and dc.contributor.translator tags
+    and adds *_parsed versions with forenames first.
+    """
     for person in doc.xpath("|".join('//dc:'+(tag) for tag in (
-                    'creator', 'contributor.translator', 'contributor.editor', 'contributor.technical_editor')),
+                    'creator', 'contributor.translator')),
                     namespaces = {'dc': str(DCNS)})[::-1]:
         if not person.text:
             continue
@@ -179,31 +191,46 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None,
     verbose: prints all output from LaTeX
     save_tex: path to save the intermediary LaTeX file to
     morefloats (old/new/none): force specific morefloats
-    cover: a cover.Cover object
+    cover: a cover.Cover factory or True for default
     flags: less-advertising,
     customizations: user requested customizations regarding various formatting parameters (passed to wl LaTeX class)
     """
 
     # Parse XSLT
     try:
+        book_info = wldoc.book_info
         document = load_including_children(wldoc)
+        root = document.edoc.getroot()
 
         if cover:
-            document.edoc.getroot().set('data-cover-width', str(cover.width))
-            document.edoc.getroot().set('data-cover-height', str(cover.height))
+            if cover is True:
+                cover = WLCover
+            bound_cover = cover(book_info)
+            root.set('data-cover-width', str(bound_cover.width))
+            root.set('data-cover-height', str(bound_cover.height))
+            if bound_cover.uses_dc_cover:
+                if book_info.cover_by:
+                    root.set('data-cover-by', book_info.cover_by)
+                if book_info.cover_source:
+                    root.set('data-cover-source',
+                            book_info.cover_source)
         if flags:
             for flag in flags:
-                document.edoc.getroot().set('flag-' + flag, 'yes')
+                root.set('flag-' + flag, 'yes')
 
         # check for LaTeX packages
         if morefloats:
-            document.edoc.getroot().set('morefloats', morefloats.lower())
+            root.set('morefloats', morefloats.lower())
         elif package_available('morefloats', 'maxfloats=19'):
-            document.edoc.getroot().set('morefloats', 'new')
+            root.set('morefloats', 'new')
 
         # add customizations
         if customizations is not None:
-            document.edoc.getroot().set('customizations', u','.join(customizations))
+            root.set('customizations', u','.join(customizations))
+
+        # add editors info
+        root.set('editors', u', '.join(sorted(
+            editor.readable() for editor in document.editors())))
 
         # hack the tree
         move_motifs_inside(document.edoc)
@@ -222,9 +249,8 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None,
         temp = mkdtemp('-wl2pdf')
 
         if cover:
-            c = cover(document.book_info.author.readable(), document.book_info.title)
             with open(os.path.join(temp, 'cover.png'), 'w') as f:
-                c.save(f)
+                bound_cover.save(f)
 
         del document # no longer needed large object :)
 
@@ -241,7 +267,10 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None,
         shutil.copy(get_resource('pdf/wl.cls'), temp)
         shutil.copy(get_resource('res/wl-logo.png'), temp)
 
-        cwd = os.getcwd()
+        try:
+            cwd = os.getcwd()
+        except OSError:
+            cwd = None
         os.chdir(temp)
 
         if verbose:
@@ -251,7 +280,8 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None,
         if p:
             raise ParseError("Error parsing .tex file")
 
-        os.chdir(cwd)
+        if cwd is not None:
+            os.chdir(cwd)
 
         output_file = NamedTemporaryFile(prefix='librarian', suffix='.pdf', delete=False)
         pdf_path = os.path.join(temp, 'doc.pdf')
@@ -281,7 +311,8 @@ def load_including_children(wldoc=None, provider=None, uri=None):
 
     text = re.sub(ur"([\u0400-\u04ff]+)", ur"<alien>\1</alien>", text)
 
-    document = WLDocument.from_string(text, parse_dublincore=True)
+    document = WLDocument.from_string(text,
+                parse_dublincore=True, provider=provider)
     document.swap_endlines()
 
     for child_uri in document.book_info.parts: