X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/5b174651e7cadaa9c353cd247fbfcde41b012f3a..890a8fb6adbf3a3d4eb2a8594baff30a9ec366b9:/librarian/epub.py

diff --git a/librarian/epub.py b/librarian/epub.py
index 139dd39..ef07c3f 100644
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -7,16 +7,21 @@ from __future__ import with_statement
 
 import os
 import os.path
-import shutil
-import sys
+import subprocess
 from copy import deepcopy
 from lxml import etree
 import zipfile
+from tempfile import mkdtemp
+from shutil import rmtree
+
+import sys
 
-from librarian import XMLNamespace, RDFNS, DCNS, WLNS, XHTMLNS, NoDublinCore
+from librarian import XMLNamespace, RDFNS, DCNS, WLNS, NCXNS, OPFNS, NoDublinCore
+from librarian.dcparser import BookInfo
 
-NCXNS = XMLNamespace("http://www.daisy.org/z3986/2005/ncx/")
-OPFNS = XMLNamespace("http://www.idpf.org/2007/opf")
+from librarian import functions
+
+functions.reg_person_name()
 
 
 def inner_xml(node):
@@ -38,7 +43,6 @@ def set_inner_xml(node, text):
     <a>x<b>y</b>z</a>
     """
 
-    
     p = etree.fromstring('<x>%s</x>' % text)
     node.text = p.text
     node[:] = p[:]
@@ -78,12 +82,11 @@ def replace_characters(node):
     def replace_chars(text):
         if text is None:
             return None
-        return text.replace("&", "&amp;")\
-                   .replace("---", "&#8212;")\
-                   .replace("--", "&#8211;")\
-                   .replace(",,", "&#8222;")\
-                   .replace('"', "&#8221;")\
-                   .replace("'", "&#8217;")
+        return text.replace("---", u"\u2014")\
+                   .replace("--", u"\u2013")\
+                   .replace(",,", u"\u201E")\
+                   .replace('"', u"\u201D")\
+                   .replace("'", u"\u2019")
     if node.tag == 'extra':
         node.clear()
     else:
@@ -93,24 +96,26 @@ def replace_characters(node):
             replace_characters(child)
 
 
-def find_annotations(annotations, source, part_number):
+def find_annotations(annotations, source, part_no):
     for child in source:
         if child.tag in ('pe', 'pa', 'pt', 'pr'):
             annotation = deepcopy(child)
-            annotation.set('number', str(len(annotations)+1))
-            annotation.set('part', str(part_number))
+            number = str(len(annotations)+1)
+            annotation.set('number', number)
+            annotation.set('part', str(part_no))
             annotation.tail = ''
             annotations.append(annotation)
             tail = child.tail
             child.clear()
             child.tail = tail
-            child.text = str(len(annotations))
+            child.text = number
         if child.tag not in ('extra', 'podtytul'):
-            find_annotations(annotations, child, part_number)
+            find_annotations(annotations, child, part_no)
 
 
 def replace_by_verse(tree):
     """ Find stanzas and create new verses in place of a '/' character """
+
     stanzas = tree.findall('.//' + WLNS('strofa'))
     for node in stanzas:
         for child_node in node:
@@ -137,6 +142,7 @@ def replace_by_verse(tree):
 
 def add_to_manifest(manifest, partno):
     """ Adds a node to the manifest section in content.opf file """
+
     partstr = 'part%d' % partno
     e = manifest.makeelement(OPFNS('item'), attrib={
                                  'id': partstr,
@@ -148,84 +154,232 @@ def add_to_manifest(manifest, partno):
 
 def add_to_spine(spine, partno):
     """ Adds a node to the spine section in content.opf file """
+
     e = spine.makeelement(OPFNS('itemref'), attrib={'idref': 'part%d' % partno});
     spine.append(e)
 
 
-def add_nav_point(nav_map, counter, title, part_counter):
-    nav_point = nav_map.makeelement(NCXNS('navPoint'))
-    nav_point.set('id', 'NavPoint-%d' % counter)
-    nav_point.set('playOrder', str(counter))
-
-    nav_label = nav_map.makeelement(NCXNS('navLabel'))
-    text = nav_map.makeelement(NCXNS('text'))
-    text.text = title
-    nav_label.append(text)
-    nav_point.append(nav_label)
-
-    content = nav_map.makeelement(NCXNS('content'))
-    content.set('src', 'part%d.html' % part_counter)
-    nav_point.append(content)
+class TOC(object):
+    def __init__(self, name=None, part_number=None):
+        self.children = []
+        self.name = name
+        self.part_number = part_number
+        self.sub_number = None
 
-    nav_map.append(nav_point)
-
-
-def add_nav_point2(nav_map, counter, title, part_counter, subcounter):
-    nav_point = nav_map.makeelement(NCXNS('navPoint'))
-    nav_point.set('id', 'NavPoint-%d' % counter)
-    nav_point.set('playOrder', str(counter))
-
-    nav_label = nav_map.makeelement(NCXNS('navLabel'))
-    text = nav_map.makeelement(NCXNS('text'))
-    text.text = title
-    nav_label.append(text)
-    nav_point.append(nav_label)
-
-    content = nav_map.makeelement(NCXNS('content'))
-    content.set('src', 'part%d.html#sub%d' % (part_counter, subcounter))
-    nav_point.append(content)
-
-    nav_map[-1].append(nav_point)
-
-
-def transform(input_file, output_file):
-    """ produces an epub
+    def add(self, name, part_number, level=0, is_part=True):
+        if level > 0 and self.children:
+            return self.children[-1].add(name, part_number, level-1, is_part)
+        else:
+            t = TOC(name)
+            t.part_number = part_number
+            self.children.append(t)
+            if not is_part:
+                t.sub_number = len(self.children) + 1
+                return t.sub_number
+
+    def append(self, toc):
+        self.children.append(toc)
+
+    def extend(self, toc):
+        self.children.extend(toc.children)
+
+    def depth(self):
+        if self.children:
+            return max((c.depth() for c in self.children)) + 1
+        else:
+            return 0
+
+    def write_to_xml(self, nav_map, counter):
+        for child in self.children:
+            nav_point = nav_map.makeelement(NCXNS('navPoint'))
+            nav_point.set('id', 'NavPoint-%d' % counter)
+            nav_point.set('playOrder', str(counter))
+
+            nav_label = nav_map.makeelement(NCXNS('navLabel'))
+            text = nav_map.makeelement(NCXNS('text'))
+            text.text = child.name
+            nav_label.append(text)
+            nav_point.append(nav_label)
+
+            content = nav_map.makeelement(NCXNS('content'))
+            src = 'part%d.html' % child.part_number
+            if child.sub_number is not None:
+                src += '#sub%d' % child.sub_number
+            content.set('src', src)
+            nav_point.append(content)
+            nav_map.append(nav_point)
+            counter = child.write_to_xml(nav_point, counter + 1)
+        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 """
+
+    # prepare a container for each chunk
+    part_xml = etree.Element('utwor')
+    etree.SubElement(part_xml, 'master')
+    main_xml_part = part_xml[0] # master
 
-    input_file and output_file should be filelike objects
+    last_node_part = False
+    for one_part in main_text:
+        name = one_part.tag
+        if name == 'naglowek_czesc':
+            yield part_xml
+            last_node_part = True
+            main_xml_part[:] = [deepcopy(one_part)]
+        elif not last_node_part and name in ("naglowek_rozdzial", "naglowek_akt", "srodtytul"):
+            yield part_xml
+            main_xml_part[:] = [deepcopy(one_part)]
+        else:
+            main_xml_part.append(deepcopy(one_part))
+            last_node_part = False
+    yield part_xml
+
+
+def transform_chunk(chunk_xml, chunk_no, annotations):
+    """ transforms one chunk, returns a HTML string, a TOC object and a set of used characters """
+
+    toc = TOC()
+    for element in chunk_xml[0]:
+        if element.tag in ("naglowek_czesc", "naglowek_rozdzial", "naglowek_akt", "srodtytul"):
+            toc.add(node_name(element), chunk_no)
+        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)
+    html_tree = xslt(chunk_xml, res('xsltScheme.xsl'))
+    chars = used_chars(html_tree.getroot())
+    output_html = etree.tostring(html_tree, 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):
+    """ 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
     """
 
-    input_xml = etree.parse(input_file)
+    def transform_file(input_xml, chunk_counter=1, first=True):
+        """ processes one input file and proceeds to its children """
+
+        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
+            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))
+        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))
+            add_to_manifest(manifest, chunk_counter)
+            add_to_spine(spine, chunk_counter)
+            chunk_counter += 1
+
+        if len(input_xml.getroot()) > 1:
+            # rdf before style master
+            main_text = input_xml.getroot()[1]
+        else:
+            # rdf in style master
+            main_text = input_xml.getroot()[0]
+            if main_text.tag == RDFNS('RDF'):
+                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)
+                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)
+                chunk_counter += 1
+
+        if children:
+            for child in children:
+                child_xml = etree.parse(provider.by_uri(child))
+                child_toc, chunk_counter, chunk_chars = transform_file(child_xml, chunk_counter, first=False)
+                toc.append(child_toc)
+                chars = chars.union(chunk_chars)
+
+        return toc, chunk_counter, chars
+
+    # 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])
+
+    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)
 
+    # write static elements
     mime = zipfile.ZipInfo()
     mime.filename = 'mimetype'
     mime.compress_type = zipfile.ZIP_STORED
     mime.extra = ''
     zip.writestr(mime, 'application/epub+zip')
-
     zip.writestr('META-INF/container.xml', '<?xml version="1.0" ?><container version="1.0" ' \
                        'xmlns="urn:oasis:names:tc:opendocument:xmlns:container">' \
                        '<rootfiles><rootfile full-path="OPS/content.opf" ' \
                        'media-type="application/oebps-package+xml" />' \
                        '</rootfiles></container>')
+    for fname in 'style.css', 'logo_wolnelektury.png':
+        zip.write(res(fname), os.path.join('OPS', fname))
 
-    metadata_el = input_xml.find('.//'+RDFNS('Description'))
-    if metadata_el is None:
-        raise NoDublinCore('Document has no DublinCore - which is required.')
-    metadatasource = etree.ElementTree(metadata_el)
-
-    opf = xslt(metadatasource, res('xsltContent.xsl'))
-
+    opf = xslt(metadata, res('xsltContent.xsl'))
     manifest = opf.find('.//' + OPFNS('manifest'))
     spine = opf.find('.//' + OPFNS('spine'))
 
-    for fname in 'style.css', 'logo_wolnelektury.png':
-        zip.write(res(fname), os.path.join('OPS', fname))
-
     annotations = etree.Element('annotations')
-    part_xml = etree.Element('utwor')
-    etree.SubElement(part_xml, 'master')
 
     toc_file = etree.fromstring('<?xml version="1.0" encoding="utf-8"?><!DOCTYPE ncx PUBLIC ' \
                                '"-//NISO//DTD ncx 2005-1//EN" "http://www.daisy.org/z3986/2005/ncx-2005-1.dtd">' \
@@ -233,99 +387,45 @@ def transform(input_file, 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" />' \
-                               '</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]
 
-    main_xml_part = part_xml[0] # było [0][0], master
-    nav_map = toc_file[-1] # było [-1][-1]
-    depth = 1 # navmap
-
-    if len(input_xml.getroot()) > 1:
-        # rdf before style master
-        main_text = input_xml.getroot()[1]
-    else:
-        # rdf in style master
-        main_text = input_xml.getroot()[0]
-
-    replace_characters(main_text)
-    zip.writestr('OPS/title.html', 
-                 etree.tostring(xslt(input_xml, res('xsltTitle.xsl')), pretty_print=True))
-
-    # Search for table of contents elements and book division
-
-    stupid_i = stupid_j = stupid_k = 1
-    last_node_part = False
-    for one_part in main_text:
-        name = one_part.tag
-        if name in ("naglowek_czesc", "naglowek_rozdzial", "naglowek_akt", "srodtytul"):
-            if name == "naglowek_czesc":
-                stupid_k = 1
-                last_node_part = True
-                find_annotations(annotations, part_xml, stupid_j)
-                replace_by_verse(part_xml)
-                zip.writestr('OPS/part%d.html' % stupid_j,
-                            etree.tostring(xslt(part_xml, res('xsltScheme.xsl')), pretty_print=True))
-                main_xml_part[:] = [deepcopy(one_part)]
-                # add to manifest and spine
-                add_to_manifest(manifest, stupid_j)
-                add_to_spine(spine, stupid_j)
-                name_toc = node_name(one_part)
-                # build table of contents
-                # i+2 because of title page
-                add_nav_point(nav_map, stupid_i+2, name_toc, stupid_j + 1)
-                stupid_i += 1
-                stupid_j += 1
-            else:
-                if last_node_part:
-                    main_xml_part.append(one_part)
-                    last_node_part = False
-                    name_toc = node_name(one_part)
-                    add_nav_point(nav_map, stupid_i + 1, name_toc, stupid_j)
-                else:
-                    stupid_k = 1
-                    find_annotations(annotations, part_xml, stupid_j)
-                    replace_by_verse(part_xml)
-                    zip.writestr('OPS/part%d.html' % stupid_j,
-                                 etree.tostring(xslt(part_xml, res('xsltScheme.xsl')), pretty_print=True))
-                    # start building a new part
-                    main_xml_part[:] = [deepcopy(one_part)]
-                    add_to_manifest(manifest, stupid_j)
-                    add_to_spine(spine, stupid_j)
-                    name_toc = node_name(one_part)
-                    add_nav_point(nav_map, stupid_i + 2, name_toc, stupid_j + 1) # title page
-                    stupid_j += 1
-                    stupid_i += 1
-        else:
-            if name in ('naglowek_podrozdzial', 'naglowek_scena'):
-                depth = 2
-                name_toc =  node_name(one_part)
-                add_nav_point2(nav_map, stupid_i + 2, name_toc, stupid_j, stupid_k)
-                one_part.set('sub', str(stupid_k))
-                stupid_k += 1
-                stupid_i += 1
-            main_xml_part.append(deepcopy(one_part))
-            last_node_part = False
-    find_annotations(annotations, part_xml, stupid_j)
-    replace_by_verse(part_xml)
-    add_to_manifest(manifest, stupid_j)
-    add_to_spine(spine, stupid_j)
+    toc, chunk_counter, chars = transform_file(input_xml)
 
-    zip.writestr('OPS/part%d.html' % stupid_j,
-                 etree.tostring(xslt(part_xml, res('xsltScheme.xsl')), pretty_print=True))
+    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:
         nav_map.append(etree.fromstring(
             '<navPoint id="NavPoint-%(i)d" playOrder="%(i)d" ><navLabel><text>Przypisy</text>'\
-            '</navLabel><content src="annotations.html" /></navPoint>' % {'i':stupid_i+2}))
+            '</navLabel><content src="annotations.html" /></navPoint>' % {'i': toc_counter}))
         manifest.append(etree.fromstring(
             '<item id="annotations" href="annotations.html" media-type="application/xhtml+xml" />'))
         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(
-                            xslt(annotations, res("xsltAnnotations.xsl")), pretty_print=True))
+                            html_tree, 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 = []
@@ -337,31 +437,7 @@ def transform(input_file, output_file):
         meta.set('content', '0')
         toc_file[0].append(meta)
     toc_file[0][0].set('content', ''.join((title, 'WolneLektury.pl')))
-    toc_file[0][1].set('content', str(depth))
+    toc_file[0][1].set('content', str(toc.depth()))
     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__':
-    import html
-
-    if len(sys.argv) < 2:
-        print >> sys.stderr, 'Usage: python epub.py <input file> [output file]'
-        sys.exit(1)
-
-    input = sys.argv[1]
-    if len(sys.argv) > 2:
-        output = sys.argv[2]
-    else:
-        basename, ext = os.path.splitext(input)
-        output = basename + '.epub' 
-
-    print input
-    if html.transform(input, is_file=True) == '<empty />':
-        print 'empty content - skipping'
-    else:
-        transform(open(input, 'r'), open(output, 'w'))
-
-
-