1 # -*- coding: utf-8 -*-
3 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
4 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
6 """PDF creation library.
8 Creates one big XML from the book and its children, converts it to LaTeX
9 with TeXML, then runs it by XeLaTeX.
12 from __future__ import with_statement
16 from distutils.dir_util import copy_tree # shutil.copytree is so uncapable.
17 from StringIO import StringIO
18 from tempfile import mkdtemp, NamedTemporaryFile
20 from copy import deepcopy
21 from subprocess import call, PIPE
23 from Texml.processor import process
24 from lxml import etree
25 from lxml.etree import XMLSyntaxError, XSLTApplyError
27 from librarian.dcparser import Person
28 from librarian.parser import WLDocument
29 from librarian import ParseError, DCNS, get_resource, OutputFile
30 from librarian import functions
31 from librarian.cover import WLCover
34 functions.reg_substitute_entities()
36 functions.reg_starts_white()
37 functions.reg_ends_white()
38 functions.reg_texcommand()
41 'wl2tex': 'pdf/wl2tex.xslt',
53 def insert_tags(doc, split_re, tagname, exclude=None):
54 """ inserts <tagname> for every occurence of `split_re' in text nodes in the `doc' tree
56 >>> t = etree.fromstring('<a><b>A-B-C</b>X-Y-Z</a>');
57 >>> insert_tags(t, re.compile('-'), 'd');
58 >>> print etree.tostring(t)
59 <a><b>A<d/>B<d/>C</b>X<d/>Y<d/>Z</a>
62 for elem in doc.iter(tag=etree.Element):
63 if exclude and elem.tag in exclude:
66 chunks = split_re.split(elem.text)
67 while len(chunks) > 1:
68 ins = etree.Element(tagname)
69 ins.tail = chunks.pop()
71 elem.text = chunks.pop(0)
73 chunks = split_re.split(elem.tail)
74 parent = elem.getparent()
75 ins_index = parent.index(elem) + 1
76 while len(chunks) > 1:
77 ins = etree.Element(tagname)
78 ins.tail = chunks.pop()
79 parent.insert(ins_index, ins)
80 elem.tail = chunks.pop(0)
83 def substitute_hyphens(doc):
85 re.compile("(?<=[^-\s])-(?=[^-\s])"),
87 exclude=[DCNS("identifier.url"), DCNS("rights.license"), "www"]
93 re.compile("(?<=\s\w)\s+"),
95 exclude=[DCNS("identifier.url"), DCNS("rights.license")]
99 def move_motifs_inside(doc):
100 """ moves motifs to be into block elements """
101 for master in doc.xpath('//powiesc|//opowiadanie|//liryka_l|//liryka_lp|//dramat_wierszowany_l|//dramat_wierszowany_lp|//dramat_wspolczesny'):
102 for motif in master.xpath('motyw'):
103 for sib in motif.itersiblings():
104 if sib.tag not in ('sekcja_swiatlo', 'sekcja_asterysk', 'separator_linia', 'begin', 'end', 'motyw', 'extra', 'uwaga'):
105 # motif shouldn't have a tail - it would be untagged text
107 motif.getparent().remove(motif)
112 def hack_motifs(doc):
113 """ dirty hack for the marginpar-creates-orphans LaTeX problem
114 see http://www.latex-project.org/cgi-bin/ltxbugs2html?pr=latex/2304
116 moves motifs in stanzas from first verse to second
117 and from next to last to last, then inserts negative vspace before them
119 for motif in doc.findall('//strofa//motyw'):
120 # find relevant verse-level tag
121 verse, stanza = motif, motif.getparent()
122 while stanza is not None and stanza.tag != 'strofa':
123 verse, stanza = stanza, stanza.getparent()
124 breaks_before = sum(1 for i in verse.itersiblings('br', preceding=True))
125 breaks_after = sum(1 for i in verse.itersiblings('br'))
126 if (breaks_before == 0 and breaks_after > 0) or breaks_after == 1:
128 if breaks_after == 2:
130 moved_motif = deepcopy(motif)
133 moved_motif.tail = None
134 moved_motif.set('moved', str(move_by))
136 for br in verse.itersiblings('br'):
140 br.addnext(moved_motif)
144 def parse_creator(doc):
145 """Generates readable versions of creator and translator tags.
147 Finds all dc:creator and dc.contributor.translator tags
148 and adds *_parsed versions with forenames first.
150 for person in doc.xpath("|".join('//dc:'+(tag) for tag in (
151 'creator', 'contributor.translator')),
152 namespaces = {'dc': str(DCNS)})[::-1]:
155 p = Person.from_text(person.text)
156 person_parsed = deepcopy(person)
157 person_parsed.tag = person.tag + '_parsed'
158 person_parsed.set('sortkey', person.text)
159 person_parsed.text = p.readable()
160 person.getparent().insert(0, person_parsed)
163 def get_stylesheet(name):
164 return get_resource(STYLESHEETS[name])
167 def package_available(package, args='', verbose=False):
168 """ check if a verion of a latex package accepting given args is available """
169 tempdir = mkdtemp('-wl2pdf-test')
170 fpath = os.path.join(tempdir, 'test.tex')
177 """ % (args, package))
180 p = call(['xelatex', '-output-directory', tempdir, fpath])
182 p = call(['xelatex', '-interaction=batchmode', '-output-directory', tempdir, fpath], stdout=PIPE, stderr=PIPE)
183 shutil.rmtree(tempdir)
187 def transform(wldoc, verbose=False, save_tex=None, save_texml=None, morefloats=None,
188 cover=None, flags=None, customizations=None, documentclass='wl', resources=None):
189 """ produces a PDF file with XeLaTeX
192 verbose: prints all output from LaTeX
193 save_tex: path to save the intermediary LaTeX file to
194 save_texml: path to save the intermediary TeXML file to
195 morefloats (old/new/none): force specific morefloats
196 cover: a cover.Cover factory or True for default
197 flags: less-advertising,
198 customizations: user requested customizations regarding various formatting parameters (passed to wl LaTeX class)
199 documentclass: LaTeX document class, defaults to wl
200 resources: a directory with resources, copied to place where LaTeX compilation is made
205 book_info = wldoc.book_info
206 document = load_including_children(wldoc)
207 root = document.edoc.getroot()
212 bound_cover = cover(book_info)
213 root.set('data-cover-width', str(bound_cover.width))
214 root.set('data-cover-height', str(bound_cover.height))
215 if bound_cover.uses_dc_cover:
216 if book_info.cover_by:
217 root.set('data-cover-by', book_info.cover_by)
218 if book_info.cover_source:
219 root.set('data-cover-source',
220 book_info.cover_source)
223 root.set('flag-' + flag, 'yes')
225 # check for LaTeX packages
227 root.set('morefloats', morefloats.lower())
228 elif package_available('morefloats', 'maxfloats=19'):
229 root.set('morefloats', 'new')
232 if customizations is not None:
233 root.set('customizations', u','.join(customizations))
235 root.set('documentclass', documentclass or 'wl')
238 root.set('editors', u', '.join(sorted(
239 editor.readable() for editor in document.editors())))
242 move_motifs_inside(document.edoc)
243 hack_motifs(document.edoc)
244 parse_creator(document.edoc)
245 substitute_hyphens(document.edoc)
246 fix_hanging(document.edoc)
249 style_filename = get_stylesheet("wl2tex")
250 style = etree.parse(style_filename)
252 texml = document.transform(style)
255 texml.write(save_texml)
258 temp = mkdtemp('-wl2pdf')
261 with open(os.path.join(temp, 'cover.png'), 'w') as f:
264 del document # no longer needed large object :)
266 tex_path = os.path.join(temp, 'doc.tex')
267 fout = open(tex_path, 'w')
268 process(StringIO(texml), fout, 'utf-8')
273 shutil.copy(tex_path, save_tex)
276 shutil.copy(get_resource('pdf/wl.cls'), temp)
277 shutil.copy(get_resource('pdf/wlpub.cls'), temp)
278 shutil.copy(get_resource('res/wl-logo.png'), temp)
280 copy_tree(resources, temp)
289 p = call(['xelatex', tex_path])
291 p = call(['xelatex', '-interaction=batchmode', tex_path], stdout=PIPE, stderr=PIPE)
293 raise ParseError("Error parsing .tex file")
298 output_file = NamedTemporaryFile(prefix='librarian', suffix='.pdf', delete=False)
299 pdf_path = os.path.join(temp, 'doc.pdf')
300 shutil.move(pdf_path, output_file.name)
302 return OutputFile.from_filename(output_file.name)
304 except (XMLSyntaxError, XSLTApplyError), e:
308 def load_including_children(wldoc=None, provider=None, uri=None):
309 """ Makes one big xml file with children inserted at end.
311 Either wldoc or provider and URI must be provided.
315 f = provider.by_uri(uri)
316 text = f.read().decode('utf-8')
318 elif wldoc is not None:
319 text = etree.tostring(wldoc.edoc, encoding=unicode)
320 provider = wldoc.provider
322 raise ValueError('Neither a WLDocument, nor provider and URI were provided.')
324 text = re.sub(ur"([\u0400-\u04ff]+)", ur"<alien>\1</alien>", text)
326 document = WLDocument.from_string(text,
327 parse_dublincore=True, provider=provider)
328 document.swap_endlines()
330 for child_uri in document.book_info.parts:
331 child = load_including_children(provider=provider, uri=child_uri)
332 document.edoc.getroot().append(child.edoc.getroot())