X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/f33381c4b0468544123ff05295033da7cc4bf61f..607ea5393e3768f8b842b613b100a2fb62074f31:/librarian/epub.py
diff --git a/librarian/epub.py b/librarian/epub.py
index 3bdd1cd..8dc11c7 100644
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -1,25 +1,26 @@
# -*- coding: utf-8 -*-
#
# This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
-# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
+# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
#
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
+from mimetypes import guess_type
-import sys
-
-from librarian import XMLNamespace, RDFNS, DCNS, WLNS, NCXNS, OPFNS, NoDublinCore
-from librarian.dcparser import BookInfo
-
-from librarian import functions
+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()
@@ -32,7 +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])
+ 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
@@ -73,11 +74,6 @@ def xslt(xml, sheet):
return xml.xslt(etree.parse(xsltf))
-_resdir = os.path.join(os.path.dirname(os.path.realpath(__file__)), 'epub')
-def res(fname):
- return os.path.join(_resdir, fname)
-
-
def replace_characters(node):
def replace_chars(text):
if text is None:
@@ -88,13 +84,14 @@ 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):
@@ -110,35 +107,78 @@ def find_annotations(annotations, source, part_no):
child.clear()
child.tail = tail
child.text = number
- if child.tag not in ('extra', 'podtytul'):
+ 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("a c c/\\nbx/\\nyc/ \\nd")
+ >>> Stanza(s).versify()
+ >>> print etree.tostring(s)
+ a c cbx/
+ ycd
+
+ """
+ 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('', foreign_verse, ''))
- 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('', verse, ''))
- set_inner_xml(node, modified_inner_xml)
+ for stanza in stanzas:
+ Stanza(stanza).versify()
def add_to_manifest(manifest, partno):
@@ -161,19 +201,23 @@ 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
@@ -190,7 +234,13 @@ 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)
@@ -203,15 +253,26 @@ 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(
+ "" %
+ (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 """
@@ -245,73 +306,108 @@ def chop(main_text):
yield part_xml
-def transform_chunk(chunk_xml, chunk_no, annotations):
+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]:
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))
- 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)
+ if empty:
+ if not _empty_html_static:
+ _empty_html_static.append(open(get_resource('epub/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, get_resource('epub/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):
+def render_latex(wldoc, prefix="latex"):
+ """
+ Renders CODE 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, 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 //.epub instead of /.epub
+ sample=n: generate sample e-book (with at least n paragraphs)
+ 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):
+ def transform_file(wldoc, chunk_counter=1, first=True, sample=None):
""" processes one input file and proceeds to its children """
- 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, res('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"TytuÅ", "title.html")
+ elif wldoc.book_info.parts:
# 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, method="html", pretty_print=True))
+ if sample is not None and sample <= 0:
+ chars = set()
+ html_string = open(get_resource('epub/emptyChunk.html')).read()
+ else:
+ 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_manifest(manifest, chunk_counter)
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
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)
+ 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)
+
toc.extend(chunk_toc)
chars = chars.union(chunk_chars)
zip.writestr('OPS/part%d.html' % chunk_counter, chunk_html)
@@ -319,47 +415,32 @@ 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 = transform_file(child_xml, chunk_counter, first=False)
- 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
+ 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])
-
- 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 = deepcopy(wldoc)
+ del wldoc
+
+ if flags:
+ for flag in flags:
+ 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
@@ -373,12 +454,63 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
'' \
'')
- for fname in 'style.css', 'logo_wolnelektury.png':
- zip.write(res(fname), os.path.join('OPS', fname))
+ 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'))
+
+ 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(
+ '