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
+from mimetypes import guess_type
-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, OutputFile
+from librarian.cover import WLCover, FutureOfCopyrightCover
+from librarian.latex import LatexFragment
from librarian import functions, get_resource
functions.reg_person_name()
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 <b>c</b> <b>c</b>/\\nb<x>x/\\ny</x>c/ \\nd</strofa>")
+ >>> Stanza(s).versify()
+ >>> print etree.tostring(s)
+ <strofa><wers_normalny>a <b>c</b> <b>c</b></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:
+ 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
+ else:
+ verse.text = (verse.text or "") + verse_text
+
+ 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):
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 flatten_image_paths(wldoc):
+ root = wldoc.edoc.getroot()
+ for node in root.findall(".//ilustr"):
+ node.attrib['src'] = os.path.basename(node.attrib['src'])
+ return wldoc
+
+def render_latex(wldoc, prefix="latex"):
+ """
+ Renders <latex>CODE</latex> as images and returns
+ (changed_wldoc, [ (epub_filepath1, latexfragment_object1), ... ]
+"""
+ root = wldoc.edoc.getroot()
+ latex_nodes = root.findall(".//latex")
+ images = []
+ for ln in latex_nodes:
+ fragment = LatexFragment(ln.text, resize=40)
+ images.append((os.path.join(prefix, fragment.filename), fragment))
+ ln.tag = "img"
+ ln.text = os.path.join(prefix, fragment.filename)
+
+ return wldoc, images
+
+
+def transform(wldoc, verbose=False,
style=None, html_toc=False,
- sample=None, cover=None, flags=None):
+ sample=None, cover=None, flags=None, resources=None,
+ intro_file=None, cover_file=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'))), 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))
# add a title page TOC entry
- toc.add(u"Strona tytułowa", "title.html")
- elif children:
+ toc.add(u"Tytuł", "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)
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
+ flatten_image_paths(wldoc)
+
if main_text is not None:
for chunk_xml in chop(main_text):
empty = False
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
'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/logo.png'), os.path.join('OPS', 'logo.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'))
+ document, latex_images = render_latex(document)
+ for image in latex_images:
+ zip.write(image[1].path, os.path.join('OPS', image[0]))
+ image[1].remove()
+
+ if resources:
+ if os.path.isdir(resources):
+ for dp, dirs, files in os.walk(resources):
+ for fname in files:
+ fpath = os.path.join(dp, fname)
+ if os.path.isfile(fpath):
+ zip.write(fpath, os.path.join('OPS', fname))
+ manifest.append(etree.fromstring(
+ '<item id="%s" href="%s" media-type="%s" />' % (os.path.splitext(fname)[0], fname, guess_type(fpath)[0])))
+
+ else:
+ print "resources path %s is not directory" % resources
+
if cover:
+ if cover is True:
+ cover = FutureOfCopyrightCover
+
cover_file = StringIO()
- c = cover(book_info.author.readable(), book_info.title)
+ c = cover(document.book_info)
c.save(cover_file)
c_name = 'cover.%s' % c.ext()
zip.writestr(os.path.join('OPS', c_name), cover_file.getvalue())
zip.writestr('OPS/cover.html', etree.tostring(
cover_tree, method="html", pretty_print=True))
+ if c.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(
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')
'</navMap></ncx>')
nav_map = toc_file[-1]
+ manifest.append(etree.fromstring(
+ '<item id="first" href="first.html" media-type="application/xhtml+xml" />'))
+ spine.append(etree.fromstring(
+ '<itemref idref="first" />'))
+ html_tree = xslt(document.edoc, get_resource('epub/xsltFirst.xsl'))
+# chars.update(used_chars(html_tree.getroot()))
+ zip.writestr('OPS/first.html', etree.tostring(
+ html_tree, method="html", pretty_print=True))
+
+ if intro_file:
+ manifest.append(etree.fromstring(
+ '<item id="intro" href="intro.html" media-type="application/xhtml+xml" />'))
+ spine.append(etree.fromstring(
+ '<itemref idref="intro" />'))
+ zip.writestr('OPS/intro.html', open(intro_file or get_resource('epub/intro.html')).read())
+
+
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"/>'))
+ guide.append(etree.fromstring('<reference href="toc.html" type="toc" title="Table of Contents"/>'))
+
+ toc, chunk_counter, chars, sample = transform_file(document, sample=sample)
- toc, chunk_counter, chars, sample = transform_file(input_xml, sample=sample)
+ toc.add("Informacje redakcyjne", "first.html", index=0)
if len(toc.children) < 2:
- toc.add(u"Początek utworu", "part1.html")
+ toc.add(u"Początek książki", "part1.html")
# Last modifications in container files and EPUB creation
if len(annotations) > 0:
zip.writestr('OPS/annotations.html', etree.tostring(
html_tree, method="html", pretty_print=True))
- toc.add("Strona redakcyjna", "last.html")
+ # toc.add("Weprzyj Wolne Lektury", "support.html")
+ # manifest.append(etree.fromstring(
+ # '<item id="support" href="support.html" media-type="application/xhtml+xml" />'))
+ # spine.append(etree.fromstring(
+ # '<itemref idref="support" />'))
+ # html_string = open(get_resource('epub/support.html')).read()
+ # chars.update(used_chars(etree.fromstring(html_string)))
+ # zip.writestr('OPS/support.html', html_string)
+
+ toc.add("Informacje redakcyjne", "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))
if not flags or not 'without-fonts' 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':
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'))
toc.write_to_xml(nav_map)
zip.writestr('OPS/toc.ncx', etree.tostring(toc_file, pretty_print=True))
zip.close()
+
+ return OutputFile.from_filename(output_file.name)
fnp / librarian.git / blobdiff