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 from __future__ import with_statement
10 from StringIO import StringIO
11 from tempfile import mkdtemp
13 from copy import deepcopy
14 from subprocess import call, PIPE
18 from Texml.processor import process
19 from lxml import etree
20 from lxml.etree import XMLSyntaxError, XSLTApplyError
22 from librarian.dcparser import Person
23 from librarian.parser import WLDocument
24 from librarian import ParseError, DCNS, get_resource
25 from librarian import functions
28 functions.reg_substitute_entities()
30 functions.reg_starts_white()
31 functions.reg_ends_white()
32 functions.reg_texcommand()
35 'wl2tex': 'pdf/wl2tex.xslt',
39 def insert_tags(doc, split_re, tagname, exclude=None):
40 """ inserts <tagname> for every occurence of `split_re' in text nodes in the `doc' tree
42 >>> t = etree.fromstring('<a><b>A-B-C</b>X-Y-Z</a>');
43 >>> insert_tags(t, re.compile('-'), 'd');
44 >>> print etree.tostring(t)
45 <a><b>A<d/>B<d/>C</b>X<d/>Y<d/>Z</a>
48 for elem in doc.iter(tag=etree.Element):
49 if exclude and elem.tag in exclude:
52 chunks = split_re.split(elem.text)
53 while len(chunks) > 1:
54 ins = etree.Element(tagname)
55 ins.tail = chunks.pop()
57 elem.text = chunks.pop(0)
59 chunks = split_re.split(elem.tail)
60 parent = elem.getparent()
61 ins_index = parent.index(elem) + 1
62 while len(chunks) > 1:
63 ins = etree.Element(tagname)
64 ins.tail = chunks.pop()
65 parent.insert(ins_index, ins)
66 elem.tail = chunks.pop(0)
69 def substitute_hyphens(doc):
71 re.compile("(?<=[^-\s])-(?=[^-\s])"),
73 exclude=[DCNS("identifier.url"), DCNS("rights.license")]
79 re.compile("(?<=\s\w)\s+"),
81 exclude=[DCNS("identifier.url"), DCNS("rights.license")]
85 def move_motifs_inside(doc):
86 """ moves motifs to be into block elements """
87 for master in doc.xpath('//powiesc|//opowiadanie|//liryka_l|//liryka_lp|//dramat_wierszowany_l|//dramat_wierszowany_lp|//dramat_wspolczesny'):
88 for motif in master.xpath('motyw'):
89 for sib in motif.itersiblings():
90 if sib.tag not in ('sekcja_swiatlo', 'sekcja_asterysk', 'separator_linia', 'begin', 'end', 'motyw', 'extra', 'uwaga'):
91 # motif shouldn't have a tail - it would be untagged text
93 motif.getparent().remove(motif)
99 """ dirty hack for the marginpar-creates-orphans LaTeX problem
100 see http://www.latex-project.org/cgi-bin/ltxbugs2html?pr=latex/2304
102 moves motifs in stanzas from first verse to second
103 and from next to last to last, then inserts negative vspace before them
105 for motif in doc.findall('//strofa//motyw'):
106 # find relevant verse-level tag
107 verse, stanza = motif, motif.getparent()
108 while stanza is not None and stanza.tag != 'strofa':
109 verse, stanza = stanza, stanza.getparent()
110 breaks_before = sum(1 for i in verse.itersiblings('br', preceding=True))
111 breaks_after = sum(1 for i in verse.itersiblings('br'))
112 if (breaks_before == 0 and breaks_after > 0) or breaks_after == 1:
114 if breaks_after == 2:
116 moved_motif = deepcopy(motif)
119 moved_motif.tail = None
120 moved_motif.set('moved', str(move_by))
122 for br in verse.itersiblings('br'):
126 br.addnext(moved_motif)
130 def parse_creator(doc):
131 """ find all dc:creator and dc.contributor tags and add *_parsed versions with forenames first """
132 for person in doc.xpath("|".join('//dc:'+(tag) for tag in (
133 'creator', 'contributor.translator', 'contributor.editor', 'contributor.technical_editor')),
134 namespaces = {'dc': str(DCNS)})[::-1]:
135 p = Person.from_text(person.text)
136 person_parsed = deepcopy(person)
137 person_parsed.tag = person.tag + '_parsed'
138 person_parsed.text = p.readable()
139 person.getparent().insert(0, person_parsed)
142 def get_stylesheet(name):
143 return get_resource(STYLESHEETS[name])
146 def package_available(package, args='', verbose=False):
147 """ check if a verion of a latex package accepting given args is available """
148 tempdir = mkdtemp('-wl2pdf-test')
149 fpath = os.path.join(tempdir, 'test.tex')
156 """ % (args, package))
159 p = call(['xelatex', '-output-directory', tempdir, fpath])
161 p = call(['xelatex', '-interaction=batchmode', '-output-directory', tempdir, fpath], stdout=PIPE, stderr=PIPE)
162 shutil.rmtree(tempdir)
166 def transform(provider, slug=None, file_path=None,
167 output_file=None, output_dir=None, make_dir=False, verbose=False, save_tex=None, morefloats=None):
168 """ produces a PDF file with XeLaTeX
170 provider: a DocProvider
171 slug: slug of file to process, available by provider
172 file_path can be provided instead of a slug
173 output_file: file-like object or path to output file
174 output_dir: path to directory to save output file to; either this or output_file must be present
175 make_dir: writes output to <output_dir>/<author>/<slug>.pdf istead of <output_dir>/<slug>.pdf
176 verbose: prints all output from LaTeX
177 save_tex: path to save the intermediary LaTeX file to
178 morefloats (old/new/none): force specific morefloats
185 raise ValueError('slug or file_path should be specified, not both')
186 document = load_including_children(provider, file_path=file_path)
189 raise ValueError('either slug or file_path should be specified')
190 document = load_including_children(provider, slug=slug)
192 # check for LaTeX packages
194 document.edoc.getroot().set('morefloats', morefloats.lower())
195 elif package_available('morefloats', 'maxfloats=19'):
196 document.edoc.getroot().set('morefloats', 'new')
199 move_motifs_inside(document.edoc)
200 hack_motifs(document.edoc)
201 parse_creator(document.edoc)
202 substitute_hyphens(document.edoc)
203 fix_hanging(document.edoc)
206 if make_dir and output_dir is not None:
207 author = unicode(document.book_info.author)
208 output_dir = os.path.join(output_dir, author)
211 style_filename = get_stylesheet("wl2tex")
212 style = etree.parse(style_filename)
213 texml = document.transform(style)
214 del document # no longer needed large object :)
217 temp = mkdtemp('-wl2pdf')
218 tex_path = os.path.join(temp, 'doc.tex')
219 fout = open(tex_path, 'w')
220 process(StringIO(texml), fout, 'utf-8')
225 shutil.copy(tex_path, save_tex)
228 shutil.copy(get_resource('pdf/wl.sty'), temp)
229 shutil.copy(get_resource('res/wl-logo.png'), temp)
235 p = call(['xelatex', tex_path])
237 p = call(['xelatex', '-interaction=batchmode', tex_path], stdout=PIPE, stderr=PIPE)
239 raise ParseError("Error parsing .tex file")
244 pdf_path = os.path.join(temp, 'doc.pdf')
245 if output_dir is not None:
247 os.makedirs(output_dir)
251 output_path = os.path.join(output_dir, '%s.pdf' % slug)
253 output_path = os.path.join(output_dir, os.path.splitext(os.path.basename(file_path))[0] + '.pdf')
254 shutil.move(pdf_path, output_path)
256 if hasattr(output_file, 'write'):
258 with open(pdf_path) as f:
259 output_file.write(f.read())
262 # path to output file
263 shutil.copy(pdf_path, output_file)
266 except (XMLSyntaxError, XSLTApplyError), e:
270 def load_including_children(provider, slug=None, uri=None, file_path=None):
271 """ makes one big xml file with children inserted at end
272 either slug or uri must be provided
276 f = provider.by_uri(uri)
280 f = open(file_path, 'r')
282 raise ValueError('Neither slug, URI nor file path provided for a book.')
284 text = f.read().decode('utf-8')
285 text = re.sub(ur"([\u0400-\u04ff]+)", ur"<alien>\1</alien>", text)
287 document = WLDocument.from_string(text, True,
288 parse_dublincore=True)
291 print document.book_info
292 for child_uri in document.book_info.parts:
294 child = load_including_children(provider, uri=child_uri)
295 document.edoc.getroot().append(child.edoc.getroot())