use dejavu serif for cyrrylic

[librarian.git] / librarian / epub.py

diff --git a/librarian/epub.py b/librarian/epub.py
index 3ee487a..a5607f7 100644 (file)
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -82,7 +82,8 @@ def replace_characters(node):
     def replace_chars(text):
         if text is None:
             return None
-        return text.replace("---", u"\u2014")\
+        return text.replace(u"\ufeff", u"")\
+                   .replace("---", u"\u2014")\
                    .replace("--", u"\u2013")\
                    .replace(",,", u"\u201E")\
                    .replace('"', u"\u201D")\
@@ -258,11 +259,11 @@ def transform_chunk(chunk_xml, chunk_no, annotations):
     replace_by_verse(chunk_xml)
     html_tree = xslt(chunk_xml, res('xsltScheme.xsl'))
     chars = used_chars(html_tree.getroot())
-    output_html = etree.tostring(html_tree, pretty_print=True)
+    output_html = etree.tostring(html_tree, method="html", pretty_print=True)
     return output_html, toc, chars
 
 
-def transform(provider, slug, output_file=None, output_dir=None, make_dir=False):
+def transform(provider, slug=None, file_path=None, output_file=None, output_dir=None, make_dir=False, verbose=False):
     """ produces a EPUB file
 
     provider: a DocProvider
@@ -275,6 +276,8 @@ def transform(provider, slug, output_file=None, output_dir=None, make_dir=False)
     def transform_file(input_xml, chunk_counter=1, first=True):
         """ processes one input file and proceeds to its children """
 
+        replace_characters(input_xml.getroot())
+
         children = [child.text for child in input_xml.findall('.//'+DCNS('relation.hasPart'))]
 
         # every input file will have a TOC entry,
@@ -286,13 +289,13 @@ def transform(provider, slug, output_file=None, output_dir=None, make_dir=False)
             html_tree = xslt(input_xml, res('xsltTitle.xsl'))
             chars = used_chars(html_tree.getroot())
             zip.writestr('OPS/title.html',
-                 etree.tostring(html_tree, pretty_print=True))
+                 etree.tostring(html_tree, method="html", pretty_print=True))
         elif children:
             # write title page for every parent
             html_tree = xslt(input_xml, res('xsltChunkTitle.xsl'))
             chars = used_chars(html_tree.getroot())
             zip.writestr('OPS/part%d.html' % chunk_counter, 
-                etree.tostring(html_tree, pretty_print=True))
+                etree.tostring(html_tree, method="html", pretty_print=True))
             add_to_manifest(manifest, chunk_counter)
             add_to_spine(spine, chunk_counter)
             chunk_counter += 1
@@ -307,8 +310,6 @@ def transform(provider, slug, output_file=None, output_dir=None, make_dir=False)
                 main_text = None
 
         if main_text is not None:
-            replace_characters(main_text)
-
             for chunk_xml in chop(main_text):
                 chunk_html, chunk_toc, chunk_chars = transform_chunk(chunk_xml, chunk_counter, annotations)
                 toc.extend(chunk_toc)
@@ -328,7 +329,17 @@ def transform(provider, slug, output_file=None, output_dir=None, make_dir=False)
         return toc, chunk_counter, chars
 
     # read metadata from the first file
-    input_xml = etree.parse(provider[slug])
+    if file_path:
+        if slug:
+            raise ValueError('slug or file_path should be specified, not both')
+        f = open(file_path, 'r')
+        input_xml = etree.parse(f)
+        f.close()
+    else:
+        if not slug:
+            raise ValueError('either slug or file_path should be specified')
+        input_xml = etree.parse(provider[slug])
+
     metadata = input_xml.find('.//'+RDFNS('Description'))
     if metadata is None:
         raise NoDublinCore('Document has no DublinCore - which is required.')
@@ -344,7 +355,10 @@ def transform(provider, slug, output_file=None, output_dir=None, make_dir=False)
                 os.makedirs(output_dir)
             except OSError:
                 pass
-        output_file = open(os.path.join(output_dir, '%s.epub' % slug), 'w')
+        if slug:
+            output_file = open(os.path.join(output_dir, '%s.epub' % slug), 'w')
+        else:
+            output_file = open(os.path.join(output_dir, os.path.splitext(os.path.basename(file_path))[0] + '.epub'), 'w')
 
     zip = zipfile.ZipFile(output_file, 'w', zipfile.ZIP_DEFLATED)
 
@@ -396,7 +410,7 @@ def transform(provider, slug, output_file=None, output_dir=None, make_dir=False)
         html_tree = xslt(annotations, res("xsltAnnotations.xsl"))
         chars = chars.union(used_chars(html_tree.getroot()))
         zip.writestr('OPS/annotations.html', etree.tostring(
-                            html_tree, pretty_print=True))
+                            html_tree, method="html", pretty_print=True))
 
     # strip fonts
     tmpdir = mkdtemp('-librarian-epub')
@@ -404,7 +418,12 @@ def transform(provider, slug, output_file=None, output_dir=None, make_dir=False)
 
     os.chdir(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'font-optimizer'))
     for fname in 'DejaVuSerif.ttf', 'DejaVuSerif-Bold.ttf', 'DejaVuSerif-Italic.ttf', 'DejaVuSerif-BoldItalic.ttf':
-        subprocess.check_call(['perl', 'subset.pl', '--chars', ''.join(chars).encode('utf-8'), res('../fonts/' + fname), os.path.join(tmpdir, fname)])
+        optimizer_call = ['perl', 'subset.pl', '--chars', ''.join(chars).encode('utf-8'), res('../fonts/' + fname), os.path.join(tmpdir, fname)]
+        if verbose:
+            print "Running font-optimizer"
+            subprocess.check_call(optimizer_call)
+        else:
+            subprocess.check_call(optimizer_call, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
         zip.write(os.path.join(tmpdir, fname), os.path.join('OPS', fname))
     rmtree(tmpdir)
     os.chdir(cwd)