X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/51b9dc0babf62ea90091eeb5bafec66e39aab910..3e1dc74fbe8a20b2cc33258fad2f74c310943424:/librarian/epub.py

diff --git a/librarian/epub.py b/librarian/epub.py
index f2987b5..0c7a16c 100644
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -7,18 +7,17 @@ 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.cover import WLCover
 
 from librarian import functions, get_resource
 
@@ -84,13 +83,14 @@ def replace_characters(node):
                    .replace(",,", u"\u201E")\
                    .replace('"', u"\u201D")\
                    .replace("'", u"\u2019")
-    if node.tag == 'extra':
+    if node.tag in ('uwaga', 'extra'):
+        t = node.tail
         node.clear()
-    else:
-        node.text = replace_chars(node.text)
-        node.tail = replace_chars(node.tail)
-        for child in node:
-            replace_characters(child)
+        node.tail = t
+    node.text = replace_chars(node.text)
+    node.tail = replace_chars(node.tail)
+    for child in node:
+        replace_characters(child)
 
 
 def find_annotations(annotations, source, part_no):
@@ -106,35 +106,78 @@ def find_annotations(annotations, source, part_no):
             child.clear()
             child.tail = tail
             child.text = number
-        if child.tag not in ('extra',):
+        if child.tag not in ('extra', 'uwaga'):
             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):
@@ -157,19 +200,23 @@ def add_to_spine(spine, partno):
 
 
 class TOC(object):
-    def __init__(self, name=None, part_number=None):
+    def __init__(self, name=None, part_href=None):
         self.children = []
         self.name = name
-        self.part_number = part_number
+        self.part_href = part_href
         self.sub_number = None
 
-    def add(self, name, part_number, level=0, is_part=True):
+    def add(self, name, part_href, level=0, is_part=True, index=None):
+        assert level == 0 or index is None
         if level > 0 and self.children:
-            return self.children[-1].add(name, part_number, level-1, is_part)
+            return self.children[-1].add(name, part_href, level-1, is_part)
         else:
             t = TOC(name)
-            t.part_number = part_number
-            self.children.append(t)
+            t.part_href = part_href
+            if index is not None:
+                self.children.insert(index, t)
+            else:
+                self.children.append(t)
             if not is_part:
                 t.sub_number = len(self.children) + 1
                 return t.sub_number
@@ -186,7 +233,13 @@ class TOC(object):
         else:
             return 0
 
-    def write_to_xml(self, nav_map, counter):
+    def href(self):
+        src = self.part_href
+        if self.sub_number is not None:
+            src += '#sub%d' % self.sub_number
+        return src
+
+    def write_to_xml(self, nav_map, counter=1):
         for child in self.children:
             nav_point = nav_map.makeelement(NCXNS('navPoint'))
             nav_point.set('id', 'NavPoint-%d' % counter)
@@ -199,15 +252,26 @@ class TOC(object):
             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)
+            content.set('src', child.href())
             nav_point.append(content)
             nav_map.append(nav_point)
             counter = child.write_to_xml(nav_point, counter + 1)
         return counter
 
+    def html_part(self, depth=0):
+        texts = []
+        for child in self.children:
+            texts.append(
+                "<div style='margin-left:%dem;'><a href='%s'>%s</a></div>" %
+                (depth, child.href(), child.name))
+            texts.append(child.html_part(depth+1))
+        return "\n".join(texts)
+
+    def html(self):
+        with open(get_resource('epub/toc.html')) as f:
+            t = unicode(f.read(), 'utf-8')
+        return t % self.html_part()
+
 
 def used_chars(element):
     """ Lists characters used in an ETree Element """
@@ -247,9 +311,9 @@ def transform_chunk(chunk_xml, chunk_no, annotations, empty=False, _empty_html_s
     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)
+            toc.add(node_name(element), "part%d.html" % chunk_no)
         elif element.tag in ('naglowek_podrozdzial', 'naglowek_scena'):
-            subnumber = toc.add(node_name(element), chunk_no, level=1, is_part=False)
+            subnumber = toc.add(node_name(element), "part%d.html" % chunk_no, level=1, is_part=False)
             element.set('sub', str(subnumber))
     if empty:
         if not _empty_html_static:
@@ -265,44 +329,40 @@ 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,
+    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'))), 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))
-        elif children:
+            # add a title page TOC entry
+            toc.add(u"Strona tytułowa", "title.html")
+        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)
@@ -310,12 +370,12 @@ 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
 
@@ -336,51 +396,32 @@ 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
@@ -394,33 +435,41 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
                        '<rootfiles><rootfile full-path="OPS/content.opf" ' \
                        'media-type="application/oebps-package+xml" />' \
                        '</rootfiles></container>')
-    zip.write(get_resource('epub/style.css'), os.path.join('OPS', 'style.css'))
     zip.write(get_resource('res/wl-logo-small.png'), os.path.join('OPS', 'logo_wolnelektury.png'))
-
-    opf = xslt(metadata, get_resource('epub/xsltContent.xsl'))
-    manifest = opf.find('.//' + OPFNS('manifest'))
-    spine = opf.find('.//' + OPFNS('spine'))
+    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'))
 
     if cover:
+        if cover is True:
+            cover = WLCover
+
         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())))
-        spine.insert(0, etree.fromstring('<itemref idref="cover" />'))
+            '<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"/>'))
-        opf.getroot().append(etree.fromstring('<guide><reference href="cover.html" type="cover" title="Okładka"/></guide>'))
+        guide.append(etree.fromstring('<reference href="cover.html" type="cover" title="Okładka"/>'))
 
 
     annotations = etree.Element('annotations')
@@ -429,23 +478,24 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
                                '"-//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>' \
-                               '<navPoint id="NavPoint-1" playOrder="1"><navLabel>' \
-                               '<text>Strona tytułowa</text></navLabel><content src="title.html" />' \
-                               '</navPoint></navMap></ncx>')
+                               '</navMap></ncx>')
     nav_map = toc_file[-1]
 
-    toc, chunk_counter, chars, sample = transform_file(input_xml, sample=sample)
+    if html_toc:
+        manifest.append(etree.fromstring(
+            '<item id="html_toc" href="toc.html" media-type="application/xhtml+xml" />'))
+        spine.append(etree.fromstring(
+            '<itemref idref="html_toc" />'))
+        guide.append(etree.fromstring('<reference href="toc.html" type="toc" title="Spis treści"/>'))
+
+    toc, chunk_counter, chars, sample = transform_file(document, sample=sample)
 
-    if not toc.children:
-        toc.add(u"Początek utworu", 1)
-    toc_counter = toc.write_to_xml(nav_map, 2)
+    if len(toc.children) < 2:
+        toc.add(u"Początek utworu", "part1.html")
 
     # 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': toc_counter}))
-        toc_counter += 1
+        toc.add("Przypisy", "annotations.html")
         manifest.append(etree.fromstring(
             '<item id="annotations" href="annotations.html" media-type="application/xhtml+xml" />'))
         spine.append(etree.fromstring(
@@ -456,38 +506,42 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
         zip.writestr('OPS/annotations.html', etree.tostring(
                             html_tree, method="html", pretty_print=True))
 
-    nav_map.append(etree.fromstring(
-        '<navPoint id="NavPoint-%(i)d" playOrder="%(i)d" ><navLabel><text>Strona redakcyjna</text>'\
-        '</navLabel><content src="last.html" /></navPoint>' % {'i': toc_counter}))
+    toc.add("Strona redakcyjna", "last.html")
     manifest.append(etree.fromstring(
         '<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))
 
-    # 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'),
-                          get_resource('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)
+    if not flags or not 'without-fonts' in flags:
+        # strip fonts
+        tmpdir = mkdtemp('-librarian-epub')
+        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':
+            optimizer_call = ['perl', 'subset.pl', '--chars', ''.join(chars).encode('utf-8'),
+                              get_resource('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))
+            manifest.append(etree.fromstring(
+                '<item id="%s" href="%s" media-type="font/ttf" />' % (fname, fname)))
+        rmtree(tmpdir)
+        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'))
@@ -497,5 +551,13 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
     toc_file[0][0].set('content', ''.join((title, 'WolneLektury.pl')))
     toc_file[0][1].set('content', str(toc.depth()))
     set_inner_xml(toc_file[1], ''.join(('<text>', title, '</text>')))
+
+    # write TOC
+    if html_toc:
+        toc.add(u"Spis treści", "toc.html", index=1)
+        zip.writestr('OPS/toc.html', toc.html().encode('utf-8'))
+    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)