new curriculum
[librarian.git] / librarian / epub.py
index bb3123d..01f5c92 100644 (file)
@@ -7,18 +7,16 @@ from __future__ import with_statement
 
 import os
 import os.path
+import re
 import subprocess
 from StringIO import StringIO
 from copy import deepcopy
 from lxml import etree
 import zipfile
-from tempfile import mkdtemp
+from tempfile import mkdtemp, NamedTemporaryFile
 from shutil import rmtree
 
-import sys
-
-from librarian import XMLNamespace, RDFNS, DCNS, WLNS, NCXNS, OPFNS, XHTMLNS, NoDublinCore
-from librarian.dcparser import BookInfo
+from librarian import RDFNS, WLNS, NCXNS, OPFNS, XHTMLNS, IOFile
 
 from librarian import functions, get_resource
 
@@ -35,6 +33,7 @@ def inner_xml(node):
     nt = node.text if node.text is not None else ''
     return ''.join([nt] + [etree.tostring(child) for child in node])
 
+
 def set_inner_xml(node, text):
     """ sets node's text and children from a string
 
@@ -111,31 +110,74 @@ def find_annotations(annotations, source, part_no):
             find_annotations(annotations, child, part_no)
 
 
+class Stanza(object):
+    """
+    Converts / verse endings into verse elements in a stanza.
+
+    Slashes may only occur directly in the stanza. Any slashes in subelements
+    will be ignored, and the subelements will be put inside verse elements.
+
+    >>> s = etree.fromstring("<strofa>a/\\nb<x>x/\\ny</x>c/ \\nd</strofa>")
+    >>> Stanza(s).versify()
+    >>> print etree.tostring(s)
+    <strofa><wers_normalny>a</wers_normalny><wers_normalny>b<x>x/
+    y</x>c</wers_normalny><wers_normalny>d</wers_normalny></strofa>
+
+    """
+    def __init__(self, stanza_elem):
+        self.stanza = stanza_elem
+        self.verses = []
+        self.open_verse = None
+
+    def versify(self):
+        self.push_text(self.stanza.text)
+        for elem in self.stanza:
+            self.push_elem(elem)
+            self.push_text(elem.tail)
+        tail = self.stanza.tail
+        self.stanza.clear()
+        self.stanza.tail = tail
+        self.stanza.extend(self.verses)
+
+    def open_normal_verse(self):
+        self.open_verse = self.stanza.makeelement("wers_normalny")
+        self.verses.append(self.open_verse)
+
+    def get_open_verse(self):
+        if self.open_verse is None:
+            self.open_normal_verse()
+        return self.open_verse
+
+    def push_text(self, text):
+        if not text or not text.strip():
+            return
+        for i, verse_text in enumerate(re.split(r"/\s*\n", text)):
+            if i:
+                self.open_normal_verse()
+            verse = self.get_open_verse()
+            if len(verse):
+                verse[-1].tail = (verse[-1].tail or "") + verse_text.strip()
+            else:
+                verse.text = (verse.text or "") + verse_text.strip()
+
+    def push_elem(self, elem):
+        if elem.tag.startswith("wers"):
+            verse = deepcopy(elem)
+            verse.tail = None
+            self.verses.append(verse)
+            self.open_verse = verse
+        else:
+            appended = deepcopy(elem)
+            appended.tail = None
+            self.get_open_verse().append(appended)
+
+
 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:
-            if child_node.tag in ('slowo_obce', 'wyroznienie'):
-                foreign_verses = inner_xml(child_node).split('/\n')
-                if len(foreign_verses) > 1:
-                    new_foreign = ''
-                    for foreign_verse in foreign_verses:
-                        if foreign_verse.startswith('<wers'):
-                            new_foreign += foreign_verse
-                        else:
-                            new_foreign += ''.join(('<wers_normalny>', foreign_verse, '</wers_normalny>'))
-                    set_inner_xml(child_node, new_foreign)
-        verses = inner_xml(node).split('/\n')
-        if len(verses) > 1:
-            modified_inner_xml = ''
-            for verse in verses:
-                if verse.startswith('<wers') or verse.startswith('<extra'):
-                    modified_inner_xml += verse
-                else:
-                    modified_inner_xml += ''.join(('<wers_normalny>', verse, '</wers_normalny>'))
-            set_inner_xml(node, modified_inner_xml)
+    for stanza in stanzas:
+        Stanza(stanza).versify()
 
 
 def add_to_manifest(manifest, partno):
@@ -153,7 +195,7 @@ 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});
+    e = spine.makeelement(OPFNS('itemref'), attrib={'idref': 'part%d' % partno})
     spine.append(e)
 
 
@@ -245,7 +287,7 @@ def chop(main_text):
     # prepare a container for each chunk
     part_xml = etree.Element('utwor')
     etree.SubElement(part_xml, 'master')
-    main_xml_part = part_xml[0] # master
+    main_xml_part = part_xml[0]  # master
 
     last_node_part = False
     for one_part in main_text:
@@ -263,8 +305,10 @@ def chop(main_text):
     yield part_xml
 
 
-def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_static=[]):
+def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_static=None):
     """ transforms one chunk, returns a HTML string, a TOC object and a set of used characters """
+    if _empty_html_static is None:
+        _empty_html_static = []
 
     toc = TOC()
     for element in chunk_xml[0]:
@@ -287,47 +331,39 @@ def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_s
     return output_html, toc, chars
 
 
-def transform(provider, slug=None, file_path=None, output_file=None, output_dir=None, make_dir=False, verbose=False,
+def transform(wldoc, verbose=False,
               style=None, html_toc=False,
               sample=None, cover=None, flags=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)
-    cover: a cover.Cover object
-    flags: less-advertising, without-fonts
+    cover: a cover.Cover factory or True for default
+    flags: less-advertising, without-fonts, working-copy
     """
 
-    def transform_file(input_xml, chunk_counter=1, first=True, sample=None):
+    def transform_file(wldoc, chunk_counter=1, first=True, sample=None):
         """ 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'))]
+        replace_characters(wldoc.edoc.getroot())
 
         # every input file will have a TOC entry,
         # pointing to starting chunk
-        toc = TOC(node_name(input_xml.find('.//'+DCNS('title'))), "part%d.html" % chunk_counter)
+        toc = TOC(wldoc.book_info.title, "part%d.html" % chunk_counter)
         chars = set()
         if first:
             # write book title page
-            html_tree = xslt(input_xml, get_resource('epub/xsltTitle.xsl'))
+            html_tree = xslt(wldoc.edoc, get_resource('epub/xsltTitle.xsl'))
             chars = used_chars(html_tree.getroot())
-            zip.writestr('OPS/title.html',
-                 etree.tostring(html_tree, method="html", pretty_print=True))
+            zip.writestr('OPS/title.html', etree.tostring(html_tree, method="html", pretty_print=True))
             # add a title page TOC entry
             toc.add(u"Strona tytułowa", "title.html")
-        elif children:
+        elif wldoc.book_info.parts:
             # write title page for every parent
             if sample is not None and sample <= 0:
                 chars = set()
                 html_string = open(get_resource('epub/emptyChunk.html')).read()
             else:
-                html_tree = xslt(input_xml, get_resource('epub/xsltChunkTitle.xsl'))
+                html_tree = xslt(wldoc.edoc, get_resource('epub/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)
@@ -335,12 +371,12 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
             add_to_spine(spine, chunk_counter)
             chunk_counter += 1
 
-        if len(input_xml.getroot()) > 1:
+        if len(wldoc.edoc.getroot()) > 1:
             # rdf before style master
-            main_text = input_xml.getroot()[1]
+            main_text = wldoc.edoc.getroot()[1]
         else:
             # rdf in style master
-            main_text = input_xml.getroot()[0]
+            main_text = wldoc.edoc.getroot()[0]
             if main_text.tag == RDFNS('RDF'):
                 main_text = None
 
@@ -361,51 +397,31 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
                 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, sample = transform_file(child_xml, chunk_counter, first=False, sample=sample)
-                toc.append(child_toc)
-                chars = chars.union(chunk_chars)
+        # for child in wldoc.parts():
+        #     child_toc, chunk_counter, chunk_chars, sample = transform_file(
+        #         child, chunk_counter, first=False, sample=sample)
+        #     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])
+    document = deepcopy(wldoc)
+    del wldoc
 
     if flags:
         for flag in flags:
-            input_xml.getroot().set(flag, 'yes')
-
-    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')
+            document.edoc.getroot().set(flag, 'yes')
+
+    # add editors info
+    # document.edoc.getroot().set('editors', u', '.join(sorted(
+    #     editor.readable() for editor in document.editors())))
 
+    opf = xslt(document.book_info.to_etree(), get_resource('epub/xsltContent.xsl'))
+    manifest = opf.find('.//' + OPFNS('manifest'))
+    guide = opf.find('.//' + OPFNS('guide'))
+    spine = opf.find('.//' + OPFNS('spine'))
+
+    output_file = NamedTemporaryFile(prefix='librarian', suffix='.epub', delete=False)
     zip = zipfile.ZipFile(output_file, 'w', zipfile.ZIP_DEFLATED)
 
     # write static elements
@@ -414,51 +430,53 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
     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>')
+    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>')
     zip.write(get_resource('res/wl-logo-small.png'), os.path.join('OPS', 'logo_wolnelektury.png'))
     zip.write(get_resource('res/jedenprocent.png'), os.path.join('OPS', 'jedenprocent.png'))
     if not style:
         style = get_resource('epub/style.css')
     zip.write(style, os.path.join('OPS', 'style.css'))
 
-    opf = xslt(metadata, get_resource('epub/xsltContent.xsl'))
-    manifest = opf.find('.//' + OPFNS('manifest'))
-    guide = opf.find('.//' + OPFNS('guide'))
-    spine = opf.find('.//' + OPFNS('spine'))
-
     if cover:
         cover_file = StringIO()
-        c = cover(book_info.author.readable(), book_info.title)
-        c.save(cover_file)
-        c_name = 'cover.%s' % c.ext()
-        zip.writestr(os.path.join('OPS', c_name), cover_file.getvalue())
+        bound_cover = cover(document.book_info)
+        bound_cover.save(cover_file)
+        cover_name = 'cover.%s' % bound_cover.ext()
+        zip.writestr(os.path.join('OPS', cover_name), cover_file.getvalue())
         del cover_file
 
         cover_tree = etree.parse(get_resource('epub/cover.html'))
-        cover_tree.find('//' + XHTMLNS('img')).set('src', c_name)
+        cover_tree.find('//' + XHTMLNS('img')).set('src', cover_name)
         zip.writestr('OPS/cover.html', etree.tostring(
                         cover_tree, method="html", pretty_print=True))
 
+        if bound_cover.uses_dc_cover:
+            if document.book_info.cover_by:
+                document.edoc.getroot().set('data-cover-by', document.book_info.cover_by)
+            if document.book_info.cover_source:
+                document.edoc.getroot().set('data-cover-source', document.book_info.cover_source)
+
         manifest.append(etree.fromstring(
             '<item id="cover" href="cover.html" media-type="application/xhtml+xml" />'))
         manifest.append(etree.fromstring(
-            '<item id="cover-image" href="%s" media-type="%s" />' % (c_name, c.mime_type())))
+            '<item id="cover-image" href="%s" media-type="%s" />' % (cover_name, bound_cover.mime_type())))
         spine.insert(0, etree.fromstring('<itemref idref="cover" linear="no" />'))
         opf.getroot()[0].append(etree.fromstring('<meta name="cover" content="cover-image"/>'))
         guide.append(etree.fromstring('<reference href="cover.html" type="cover" title="Okładka"/>'))
 
-
     annotations = etree.Element('annotations')
 
-    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">' \
-                               '<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" xml:lang="pl" ' \
-                               'version="2005-1"><head></head><docTitle></docTitle><navMap>' \
-                               '</navMap></ncx>')
+    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">'
+        '<ncx xmlns="http://www.daisy.org/z3986/2005/ncx/" xml:lang="pl" '
+        'version="2005-1"><head></head><docTitle></docTitle><navMap>'
+        '</navMap></ncx>')
     nav_map = toc_file[-1]
 
     if html_toc:
@@ -468,7 +486,7 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
             '<itemref idref="html_toc" />'))
         guide.append(etree.fromstring('<reference href="toc.html" type="toc" title="Spis treści"/>'))
 
-    toc, chunk_counter, chars, sample = transform_file(input_xml, sample=sample)
+    toc, chunk_counter, chars, sample = transform_file(document, sample=sample)
 
     if len(toc.children) < 2:
         toc.add(u"Początek utworu", "part1.html")
@@ -491,15 +509,18 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
         '<item id="last" href="last.html" media-type="application/xhtml+xml" />'))
     spine.append(etree.fromstring(
         '<itemref idref="last" />'))
-    html_tree = xslt(input_xml, get_resource('epub/xsltLast.xsl'))
+    html_tree = xslt(document.edoc, get_resource('epub/xsltLast.xsl'))
     chars.update(used_chars(html_tree.getroot()))
     zip.writestr('OPS/last.html', etree.tostring(
                         html_tree, method="html", pretty_print=True))
 
-    if not flags or not 'without-fonts' in flags:
+    if not flags or 'without-fonts' not in flags:
         # strip fonts
         tmpdir = mkdtemp('-librarian-epub')
-        cwd = os.getcwd()
+        try:
+            cwd = os.getcwd()
+        except OSError:
+            cwd = None
 
         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':
@@ -514,11 +535,11 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
             manifest.append(etree.fromstring(
                 '<item id="%s" href="%s" media-type="font/ttf" />' % (fname, fname)))
         rmtree(tmpdir)
-        os.chdir(cwd)
+        if cwd is not None:
+            os.chdir(cwd)
 
     zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True))
-    contents = []
-    title = node_name(etree.ETXPath('.//'+DCNS('title'))(input_xml)[0])
+    title = document.book_info.title
     attributes = "dtb:uid", "dtb:depth", "dtb:totalPageCount", "dtb:maxPageNumber"
     for st in attributes:
         meta = toc_file.makeelement(NCXNS('meta'))
@@ -536,3 +557,5 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
     toc.write_to_xml(nav_map)
     zip.writestr('OPS/toc.ncx', etree.tostring(toc_file, pretty_print=True))
     zip.close()
+
+    return IOFile.from_filename(output_file.name)