#1032: epubs for virtualo

[librarian.git] / librarian / epub.py

diff --git a/librarian/epub.py b/librarian/epub.py
index 92df2d4..2fec331 100644 (file)
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -7,41 +7,21 @@ from __future__ import with_statement
 
 import os
 import os.path
 
 import os
 import os.path

-import shutil
-import sys
+import subprocess

 from copy import deepcopy
 from lxml import etree
 import zipfile
 from copy import deepcopy
 from lxml import etree
 import zipfile

+from tempfile import mkdtemp
+from shutil import rmtree

 
 

-from librarian import XMLNamespace, RDFNS, DCNS, WLNS, XHTMLNS, NoDublinCore
-from librarian.parser import WLDocument
-
-#TODO: shouldn't be repeated here
-NCXNS = XMLNamespace("http://www.daisy.org/z3986/2005/ncx/")
-OPFNS = XMLNamespace("http://www.idpf.org/2007/opf")
-
-
-class DocProvider(object):
-    class DoesNotExist(Exception):
-        pass
-    
-    def by_slug(self, slug):
-        raise NotImplemented
-
-    def __getitem__(self, slug):
-        return self.by_slug(slug)
-
-    def by_uri(self, uri):
-        return self.by_slug(uri.rsplit('/', 1)[1])
+import sys

 
 

+from librarian import XMLNamespace, RDFNS, DCNS, WLNS, NCXNS, OPFNS, NoDublinCore
+from librarian.dcparser import BookInfo

 
 

-class DirDocProvider(DocProvider):
-    def __init__(self, dir):
-        self.dir = dir
-        self.files = {}
+from librarian import functions

 
 

-    def by_slug(self, slug):
-        return open(os.path.join(self.dir, '%s.xml' % slug))
+functions.reg_person_name()

 
 
 def inner_xml(node):
 
 
 def inner_xml(node):


@@ -63,7 +43,6 @@ def set_inner_xml(node, text):
     <a>x<b>y</b>z</a>
     """
 
     <a>x<b>y</b>z</a>
     """
 

-    

     p = etree.fromstring('<x>%s</x>' % text)
     node.text = p.text
     node[:] = p[:]
     p = etree.fromstring('<x>%s</x>' % text)
     node.text = p.text
     node[:] = p[:]


@@ -103,7 +82,8 @@ def replace_characters(node):
     def replace_chars(text):
         if text is None:
             return None
     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")\
                    .replace("--", u"\u2013")\
                    .replace(",,", u"\u201E")\
                    .replace('"', u"\u201D")\


@@ -233,6 +213,14 @@ class TOC(object):
         return counter
 
 
         return counter
 
 

+def used_chars(element):
+    """ Lists characters used in an ETree Element """
+    chars = set((element.text or '') + (element.tail or ''))
+    for child in element:
+        chars = chars.union(used_chars(child))
+    return chars
+
+

 def chop(main_text):
     """ divide main content of the XML file into chunks """
 
 def chop(main_text):
     """ divide main content of the XML file into chunks """
 


@@ -257,8 +245,8 @@ def chop(main_text):
     yield part_xml
 
 
     yield part_xml
 
 

-def transform_chunk(chunk_xml, chunk_no, annotations):
-    """ transforms one chunk, returns a HTML string and a TOC object """
+def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_static=[]):
+    """ transforms one chunk, returns a HTML string, a TOC object and a set of used characters """

 
     toc = TOC()
     for element in chunk_xml[0]:
 
     toc = TOC()
     for element in chunk_xml[0]:


@@ -267,35 +255,58 @@ def transform_chunk(chunk_xml, chunk_no, annotations):
         elif element.tag in ('naglowek_podrozdzial', 'naglowek_scena'):
             subnumber = toc.add(node_name(element), chunk_no, level=1, is_part=False)
             element.set('sub', str(subnumber))
         elif element.tag in ('naglowek_podrozdzial', 'naglowek_scena'):
             subnumber = toc.add(node_name(element), chunk_no, level=1, is_part=False)
             element.set('sub', str(subnumber))

-    find_annotations(annotations, chunk_xml, chunk_no)
-    replace_by_verse(chunk_xml)
-    output_html = etree.tostring(xslt(chunk_xml, res('xsltScheme.xsl')), pretty_print=True)
-    return output_html, toc
-
-
-def transform(provider, slug, output_file):
-    """ produces an epub
-
-    provider is a DocProvider
-    output_file should be filelike object
+    if empty:
+        if not _empty_html_static:
+            _empty_html_static.append(open(res('emptyChunk.html')).read())
+        chars = set()
+        output_html = _empty_html_static[0]
+    else:
+        find_annotations(annotations, chunk_xml, chunk_no)
+        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, method="html", pretty_print=True)
+    return output_html, toc, chars
+
+
+def transform(provider, slug=None, file_path=None, output_file=None, output_dir=None, make_dir=False, verbose=False, sample=None):
+    """ produces a EPUB file
+
+    provider: a DocProvider
+    slug: slug of file to process, available by provider
+    output_file: file-like object or path to output file
+    output_dir: path to directory to save output file to; either this or output_file must be present
+    make_dir: writes output to <output_dir>/<author>/<slug>.epub instead of <output_dir>/<slug>.epub
+    sample=n: generate sample e-book (with at least n paragraphs)

     """
 
     """
 

-    def transform_file(input_xml, chunk_counter=1, first=True):
+    def transform_file(input_xml, chunk_counter=1, first=True, sample=None):

         """ processes one input file and proceeds to its children """
 
         """ 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,
         # pointing to starting chunk
         toc = TOC(node_name(input_xml.find('.//'+DCNS('title'))), chunk_counter)
         children = [child.text for child in input_xml.findall('.//'+DCNS('relation.hasPart'))]
 
         # every input file will have a TOC entry,
         # pointing to starting chunk
         toc = TOC(node_name(input_xml.find('.//'+DCNS('title'))), chunk_counter)

+        chars = set()

         if first:
             # write book title page
         if first:
             # write book title page

+            html_tree = xslt(input_xml, res('xsltTitle.xsl'))
+            chars = used_chars(html_tree.getroot())

             zip.writestr('OPS/title.html',
             zip.writestr('OPS/title.html',

-                 etree.tostring(xslt(input_xml, res('xsltTitle.xsl')), pretty_print=True))
+                 etree.tostring(html_tree, method="html", pretty_print=True))

         elif children:
             # write title page for every parent
         elif children:
             # write title page for every parent

-            zip.writestr('OPS/part%d.html' % chunk_counter, 
-                etree.tostring(xslt(input_xml, res('xsltChunkTitle.xsl')), pretty_print=True))
+            if sample is not None and sample <= 0:
+                chars = set()
+                html_string = open(res('emptyChunk.html')).read()
+            else:
+                html_tree = xslt(input_xml, res('xsltChunkTitle.xsl'))
+                chars = used_chars(html_tree.getroot())
+                html_string = etree.tostring(html_tree, method="html", pretty_print=True)
+            zip.writestr('OPS/part%d.html' % chunk_counter, html_string)

             add_to_manifest(manifest, chunk_counter)
             add_to_spine(spine, chunk_counter)
             chunk_counter += 1
             add_to_manifest(manifest, chunk_counter)
             add_to_spine(spine, chunk_counter)
             chunk_counter += 1


@@ -310,11 +321,17 @@ def transform(provider, slug, output_file):
                 main_text = None
 
         if main_text is not None:
                 main_text = None
 
         if main_text is not None:

-            replace_characters(main_text)
+            for chunk_xml in chop(main_text):
+                empty = False
+                if sample is not None:
+                    if sample <= 0:
+                        empty = True
+                    else:
+                        sample -= len(chunk_xml.xpath('//strofa|//akap|//akap_cd|//akap_dialog'))
+                chunk_html, chunk_toc, chunk_chars = transform_chunk(chunk_xml, chunk_counter, annotations, empty)

 
 

-            for chunk_no, chunk_xml in enumerate(chop(main_text), chunk_counter):
-                chunk_html, chunk_toc = transform_chunk(chunk_xml, chunk_counter, annotations)

                 toc.extend(chunk_toc)
                 toc.extend(chunk_toc)

+                chars = chars.union(chunk_chars)

                 zip.writestr('OPS/part%d.html' % chunk_counter, chunk_html)
                 add_to_manifest(manifest, chunk_counter)
                 add_to_spine(spine, chunk_counter)
                 zip.writestr('OPS/part%d.html' % chunk_counter, chunk_html)
                 add_to_manifest(manifest, chunk_counter)
                 add_to_spine(spine, chunk_counter)


@@ -323,11 +340,43 @@ def transform(provider, slug, output_file):
         if children:
             for child in children:
                 child_xml = etree.parse(provider.by_uri(child))
         if children:
             for child in children:
                 child_xml = etree.parse(provider.by_uri(child))

-                child_toc, chunk_counter = transform_file(child_xml, chunk_counter, first=False)
+                child_toc, chunk_counter, chunk_chars, sample = transform_file(child_xml, chunk_counter, first=False, sample=sample)

                 toc.append(child_toc)
                 toc.append(child_toc)

+                chars = chars.union(chunk_chars)
+
+        return toc, chunk_counter, chars, sample
+
+    # read metadata from the first file
+    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])

 
 

-        return toc, chunk_counter
+    metadata = input_xml.find('.//'+RDFNS('Description'))
+    if metadata is None:
+        raise NoDublinCore('Document has no DublinCore - which is required.')
+    book_info = BookInfo.from_element(input_xml)
+    metadata = etree.ElementTree(metadata)

 
 

+    # if output to dir, create the file
+    if output_dir is not None:
+        if make_dir:
+            author = unicode(book_info.author)
+            output_dir = os.path.join(output_dir, author)
+            try:
+                os.makedirs(output_dir)
+            except OSError:
+                pass
+        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)
 
 
     zip = zipfile.ZipFile(output_file, 'w', zipfile.ZIP_DEFLATED)
 


@@ -345,12 +394,6 @@ def transform(provider, slug, output_file):
     for fname in 'style.css', 'logo_wolnelektury.png':
         zip.write(res(fname), os.path.join('OPS', fname))
 
     for fname in 'style.css', 'logo_wolnelektury.png':
         zip.write(res(fname), os.path.join('OPS', fname))
 

-    # metadata from first file
-    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.')
-    metadata = etree.ElementTree(metadata)

     opf = xslt(metadata, res('xsltContent.xsl'))
     manifest = opf.find('.//' + OPFNS('manifest'))
     spine = opf.find('.//' + OPFNS('spine'))
     opf = xslt(metadata, res('xsltContent.xsl'))
     manifest = opf.find('.//' + OPFNS('manifest'))
     spine = opf.find('.//' + OPFNS('spine'))


@@ -363,13 +406,14 @@ def transform(provider, slug, output_file):
                                'version="2005-1"><head></head><docTitle></docTitle><navMap>' \
                                '<navPoint id="NavPoint-1" playOrder="1"><navLabel>' \
                                '<text>Strona tytułowa</text></navLabel><content src="title.html" />' \
                                'version="2005-1"><head></head><docTitle></docTitle><navMap>' \
                                '<navPoint id="NavPoint-1" playOrder="1"><navLabel>' \
                                '<text>Strona tytułowa</text></navLabel><content src="title.html" />' \

-                               '</navPoint><navPoint id="NavPoint-2" playOrder="2"><navLabel>' \
-                               '<text>Początek utworu</text></navLabel><content src="part1.html" />' \

                                '</navPoint></navMap></ncx>')
     nav_map = toc_file[-1]
 
                                '</navPoint></navMap></ncx>')
     nav_map = toc_file[-1]
 

-    toc, chunk_counter = transform_file(input_xml)
-    toc_counter = toc.write_to_xml(nav_map, 3) # we already have 2 navpoints
+    toc, chunk_counter, chars, sample = transform_file(input_xml, sample=sample)
+
+    if not toc.children:
+        toc.add(u"Początek utworu", 1)
+    toc_counter = toc.write_to_xml(nav_map, 2)

 
     # Last modifications in container files and EPUB creation
     if len(annotations) > 0:
 
     # Last modifications in container files and EPUB creation
     if len(annotations) > 0:


@@ -381,8 +425,26 @@ def transform(provider, slug, output_file):
         spine.append(etree.fromstring(
             '<itemref idref="annotations" />'))
         replace_by_verse(annotations)
         spine.append(etree.fromstring(
             '<itemref idref="annotations" />'))
         replace_by_verse(annotations)

+        html_tree = xslt(annotations, res("xsltAnnotations.xsl"))
+        chars = chars.union(used_chars(html_tree.getroot()))

         zip.writestr('OPS/annotations.html', etree.tostring(
         zip.writestr('OPS/annotations.html', etree.tostring(

-                            xslt(annotations, res("xsltAnnotations.xsl")), pretty_print=True))
+                            html_tree, method="html", pretty_print=True))
+
+    # strip fonts
+    tmpdir = mkdtemp('-librarian-epub')
+    cwd = os.getcwd()
+
+    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':
+        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)

 
     zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True))
     contents = []
 
     zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True))
     contents = []


@@ -398,17 +460,3 @@ def transform(provider, slug, output_file):
     set_inner_xml(toc_file[1], ''.join(('<text>', title, '</text>')))
     zip.writestr('OPS/toc.ncx', etree.tostring(toc_file, pretty_print=True))
     zip.close()
     set_inner_xml(toc_file[1], ''.join(('<text>', title, '</text>')))
     zip.writestr('OPS/toc.ncx', etree.tostring(toc_file, pretty_print=True))
     zip.close()

-
-
-if __name__ == '__main__':
-    if len(sys.argv) < 2:
-        print >> sys.stderr, 'Usage: python epub.py <input file>'
-        sys.exit(1)
-
-    main_input = sys.argv[1]
-    basepath, ext = os.path.splitext(main_input)
-    path, slug = os.path.realpath(basepath).rsplit('/', 1)
-    output = basepath + '.epub'
-    provider = DirDocProvider(path)
-    transform(provider, slug, open(output, 'w'))
-