X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/d1a6d405c4b90823231d2d12365cc133fc0edfa8..0fc9ccbbbdb5b9e714a1598ea3f201c8f6fb0a3a:/librarian/epub.py
diff --git a/librarian/epub.py b/librarian/epub.py
index c5942a2..6a2b3c5 100644
--- a/librarian/epub.py
+++ b/librarian/epub.py
@@ -7,22 +7,22 @@ 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, OutputFile
+from librarian.cover import DefaultEbookCover
from librarian import functions, get_resource
functions.reg_person_name()
+functions.reg_lang_code_3to2()
def inner_xml(node):
@@ -111,31 +111,74 @@ 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("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):
@@ -167,7 +210,7 @@ class TOC(object):
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%d.html" % part_href, level-1, is_part)
+ return self.children[-1].add(name, part_href, level-1, is_part)
else:
t = TOC(name)
t.part_href = part_href
@@ -248,18 +291,41 @@ def chop(main_text):
main_xml_part = part_xml[0] # master
last_node_part = False
+
+ # the below loop are workaround for a problem with epubs in drama ebooks without acts
+ is_scene = False
+ is_act = False
+ for one_part in main_text:
+ name = one_part.tag
+ if name == 'naglowek_scena':
+ is_scene = True
+ elif name == 'naglowek_akt':
+ is_act = True
+
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)]
+ if is_act is False and is_scene is True:
+ 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_scena"):
+ yield part_xml
+ main_xml_part[:] = [deepcopy(one_part)]
+ else:
+ main_xml_part.append(deepcopy(one_part))
+ last_node_part = False
else:
- main_xml_part.append(deepcopy(one_part))
- last_node_part = False
+ 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
@@ -268,7 +334,9 @@ 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"):
+ if element.tag in ("naglowek_czesc"):
+ toc.add(node_name(element), "part%d.html#book-text" % chunk_no)
+ elif element.tag in ("naglowek_rozdzial", "naglowek_akt", "srodtytul"):
toc.add(node_name(element), "part%d.html" % chunk_no)
elif element.tag in ('naglowek_podrozdzial', 'naglowek_scena'):
subnumber = toc.add(node_name(element), "part%d.html" % chunk_no, level=1, is_part=False)
@@ -287,47 +355,40 @@ 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 //.epub instead of /.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, with-full-fonts
"""
- 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 tytuÅowa", "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 +396,12 @@ 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 +422,37 @@ 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())))
+ if document.book_info.funders:
+ document.edoc.getroot().set('funders', u', '.join(
+ document.book_info.funders))
+ if document.book_info.thanks:
+ document.edoc.getroot().set('thanks', document.book_info.thanks)
+
+ 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
@@ -425,28 +472,32 @@ def transform(provider, slug=None, file_path=None, output_file=None, output_dir=
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:
+ if cover is True:
+ cover = DefaultEbookCover
+
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(
'