X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/167d19bdfe76a77754d2e14a80ac0c25c725c379..9a04110484aaeb16d6b6865f86039aba7e5540e5:/librarian/pdf.py diff --git a/librarian/pdf.py b/librarian/pdf.py index 659390d..03879fe 100644 --- a/librarian/pdf.py +++ b/librarian/pdf.py @@ -3,40 +3,44 @@ # This file is part of Librarian, licensed under GNU Affero GPLv3 or later. # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information. # +"""PDF creation library. + +Creates one big XML from the book and its children, converts it to LaTeX +with TeXML, then runs it by XeLaTeX. + +""" from __future__ import with_statement import os import os.path import shutil from StringIO import StringIO -from tempfile import mkdtemp +from tempfile import mkdtemp, NamedTemporaryFile import re - -import sys -sys.path.append('..') # for running from working copy +from copy import deepcopy +from subprocess import call, PIPE from Texml.processor import process from lxml import etree from lxml.etree import XMLSyntaxError, XSLTApplyError +from librarian.dcparser import Person from librarian.parser import WLDocument -from librarian import ParseError +from librarian import ParseError, DCNS, get_resource, IOFile, Format from librarian import functions - functions.reg_substitute_entities() -functions.reg_person_name() functions.reg_strip() functions.reg_starts_white() functions.reg_ends_white() +functions.reg_texcommand() STYLESHEETS = { - 'wl2tex': 'xslt/wl2tex.xslt', + 'wl2tex': 'pdf/wl2tex.xslt', } - -def insert_tags(doc, split_re, tagname): - """ inserts for every occurence of `split_re' in text nodes in the `doc' tree +def insert_tags(doc, split_re, tagname, exclude=None): + """ inserts for every occurence of `split_re' in text nodes in the `doc' tree >>> t = etree.fromstring('A-B-CX-Y-Z'); >>> insert_tags(t, re.compile('-'), 'd'); @@ -44,134 +48,271 @@ def insert_tags(doc, split_re, tagname): ABCXYZ """ - for elem in doc.iter(): + for elem in doc.iter(tag=etree.Element): + if exclude and elem.tag in exclude: + continue if elem.text: chunks = split_re.split(elem.text) - elem.text = chunks.pop(0) - while chunks: + while len(chunks) > 1: ins = etree.Element(tagname) ins.tail = chunks.pop() elem.insert(0, ins) + elem.text = chunks.pop(0) if elem.tail: chunks = split_re.split(elem.tail) parent = elem.getparent() ins_index = parent.index(elem) + 1 - elem.tail = chunks.pop(0) - while chunks: + while len(chunks) > 1: ins = etree.Element(tagname) ins.tail = chunks.pop() parent.insert(ins_index, ins) + elem.tail = chunks.pop(0) def substitute_hyphens(doc): - insert_tags(doc, + insert_tags(doc, re.compile("(?<=[^-\s])-(?=[^-\s])"), - "dywiz") + "dywiz", + exclude=[DCNS("identifier.url"), DCNS("rights.license"), 'www'] + ) def fix_hanging(doc): - insert_tags(doc, + insert_tags(doc, re.compile("(?<=\s\w)\s+"), - "nbsp") - + "nbsp", + exclude=[DCNS("identifier.url"), DCNS("rights.license")] + ) + + +def move_motifs_inside(doc): + """ moves motifs to be into block elements """ + for master in doc.xpath('//powiesc|//opowiadanie|//liryka_l|//liryka_lp|//dramat_wierszowany_l|//dramat_wierszowany_lp|//dramat_wspolczesny'): + for motif in master.xpath('motyw'): + for sib in motif.itersiblings(): + if sib.tag not in ('sekcja_swiatlo', 'sekcja_asterysk', 'separator_linia', 'begin', 'end', 'motyw', 'extra', 'uwaga'): + # motif shouldn't have a tail - it would be untagged text + motif.tail = None + motif.getparent().remove(motif) + sib.insert(0, motif) + break + + +def hack_motifs(doc): + """ dirty hack for the marginpar-creates-orphans LaTeX problem + see http://www.latex-project.org/cgi-bin/ltxbugs2html?pr=latex/2304 + + moves motifs in stanzas from first verse to second + and from next to last to last, then inserts negative vspace before them + """ + for motif in doc.findall('//strofa//motyw'): + # find relevant verse-level tag + verse, stanza = motif, motif.getparent() + while stanza is not None and stanza.tag != 'strofa': + verse, stanza = stanza, stanza.getparent() + breaks_before = sum(1 for i in verse.itersiblings('br', preceding=True)) + breaks_after = sum(1 for i in verse.itersiblings('br')) + if (breaks_before == 0 and breaks_after > 0) or breaks_after == 1: + move_by = 1 + if breaks_after == 2: + move_by += 1 + moved_motif = deepcopy(motif) + motif.tag = 'span' + motif.text = None + moved_motif.tail = None + moved_motif.set('moved', str(move_by)) + + for br in verse.itersiblings('br'): + if move_by > 1: + move_by -= 1 + continue + br.addnext(moved_motif) + break + + +def parse_creator(doc): + """Generates readable versions of creator and translator tags. + + Finds all dc:creator and dc.contributor.translator tags + and adds *_parsed versions with forenames first. + """ + for person in doc.xpath("|".join('//dc:'+(tag) for tag in ( + 'creator', 'contributor.translator')), + namespaces = {'dc': str(DCNS)})[::-1]: + if not person.text: + continue + p = Person.from_text(person.text) + person_parsed = deepcopy(person) + person_parsed.tag = person.tag + '_parsed' + person_parsed.set('sortkey', person.text) + person_parsed.text = p.readable() + person.getparent().insert(0, person_parsed) -def get_resource(path): - return os.path.join(os.path.dirname(__file__), path) def get_stylesheet(name): return get_resource(STYLESHEETS[name]) -def transform(provider, slug, output_file=None, output_dir=None): - """ produces a pdf file - provider is a DocProvider - either output_file (a file-like object) or output_dir (path to file/dir) should be specified - if output_dir is specified, file will be written to //.pdf - """ +def package_available(package, args='', verbose=False): + """ check if a verion of a latex package accepting given args is available """ + tempdir = mkdtemp('-wl2pdf-test') + fpath = os.path.join(tempdir, 'test.tex') + f = open(fpath, 'w') + f.write(r""" + \documentclass{wl} + \usepackage[%s]{%s} + \begin{document} + \end{document} + """ % (args, package)) + f.close() + if verbose: + p = call(['xelatex', '-output-directory', tempdir, fpath]) + else: + p = call(['xelatex', '-interaction=batchmode', '-output-directory', tempdir, fpath], stdout=PIPE, stderr=PIPE) + shutil.rmtree(tempdir) + return p == 0 - # Parse XSLT - try: - style_filename = get_stylesheet("wl2tex") - style = etree.parse(style_filename) - document = load_including_children(provider, slug) +def load_including_children(wldoc=None, provider=None, uri=None): + """ Makes one big xml file with children inserted at end. - substitute_hyphens(document.edoc) - fix_hanging(document.edoc) + Either wldoc or provider and URI must be provided. + """ - # if output to dir, create the file - if output_dir is not None: - author = unicode(document.book_info.author) - output_dir = os.path.join(output_dir, author) + if uri and provider: + f = provider.by_uri(uri) + text = f.read().decode('utf-8') + f.close() + elif wldoc is not None: + text = etree.tostring(wldoc.edoc, encoding=unicode) + provider = wldoc.provider + else: + raise ValueError('Neither a WLDocument, nor provider and URI were provided.') - texml = document.transform(style) - del document # no longer needed large object :) + text = re.sub(ur"([\u0400-\u04ff]+)", ur"\1", text) - temp = mkdtemp('wl2pdf-') - tex_path = os.path.join(temp, 'doc.tex') - fout = open(tex_path, 'w') - process(StringIO(texml), fout, 'utf8', 255, 0, 0) - fout.close() - del texml + document = WLDocument.from_string(text, + parse_dublincore=True, provider=provider) + document.swap_endlines() - shutil.copy(get_resource('pdf/wl.sty'), temp) - shutil.copy(get_resource('pdf/wl-logo.png'), temp) - print "pdflatex -output-directory %s %s" % (temp, os.path.join(temp, 'doc.tex')) - if os.system("pdflatex -output-directory %s %s" % (temp, os.path.join(temp, 'doc.tex'))): - raise ParseError("Error parsing .tex file") + for child_uri in document.book_info.parts: + child = load_including_children(provider=provider, uri=child_uri) + document.edoc.getroot().append(child.edoc.getroot()) + return document - pdf_path = os.path.join(temp, 'doc.pdf') - if output_dir is not None: - try: - os.makedirs(output_dir) - except OSError: - pass - output_path = os.path.join(output_dir, '%s.pdf' % slug) - shutil.move(pdf_path, output_path) - else: - with open(pdf_path) as f: - output_file.write(f.read()) - output_file.close() - return True - except (XMLSyntaxError, XSLTApplyError), e: - raise ParseError(e) +class PDFFormat(Format): + """ Base PDF format. + Available customization: + nofootnotes: Doesn't do footnotes. + nothemes: Doesn't do themes. + defaultleading: Default leading. + onehalfleading: Bigger leading. + doubleleading: Big leading. + nowlfont: Uses standard TeX font instead of JUnicodeWL. -def load_including_children(provider, slug=None, uri=None): - """ makes one big xml file with children inserted at end - either slug or uri must be provided """ - if uri: - f = provider.by_uri(uri) - elif slug: - f = provider[slug] - else: - raise ValueError('Neither slug nor URI provided for a book.') - - document = WLDocument.from_file(f, True, - parse_dublincore=True, - preserve_lines=False) + cover_class = None + tex_passes = 1 + style = get_resource('pdf/default.sty') + cover = None + + @property + def has_cover(self): + """ For use in XSLT. """ + return self.cover is not None + + @property + def customization_str(self): + """ For use in XSLT. """ + return u','.join(k for k, v in self.customization.items() if v) + + def get_document(self): + document = load_including_children(self.wldoc) + root = document.edoc.getroot() + root.set('editors', u', '.join(sorted( + editor.readable() for editor in document.editors()))) + + # hack the tree + move_motifs_inside(document.edoc) + hack_motifs(document.edoc) + parse_creator(document.edoc) + substitute_hyphens(document.edoc) + fix_hanging(document.edoc) + return document - for child_uri in document.book_info.parts: - child = load_including_children(provider, uri=child_uri) - document.edoc.getroot().append(child.edoc.getroot()) + def get_texml(self): + style_filename = get_stylesheet("wl2tex") + functions.reg_get(self) + try: + style = etree.parse(style_filename) + texml = self.get_document().transform(style) + return texml + except (XMLSyntaxError, XSLTApplyError), e: + raise ParseError(e) + + def get_tex_dir(self): + texml = self.get_texml() + temp = mkdtemp('-wl2pdf') + # Save TeX file + tex_path = os.path.join(temp, 'doc.tex') + with open(tex_path, 'w') as fout: + process(StringIO(texml), fout, 'utf-8') + if self.save_tex: + shutil.copy(tex_path, self.save_tex) + # Copy style + shutil.copy(get_resource('pdf/wl.cls'), temp) + shutil.copy(self.style, os.path.join(temp, 'style.sty')) + #for sfile in ['wasysym.sty', 'uwasyvar.fd', 'uwasy.fd']: + # shutil.copy(get_resource(os.path.join('res/wasysym', sfile)), temp) + + # Save attachments + if self.cover: + self.cover.for_pdf().dump_to(os.path.join(temp, 'makecover.sty')) + return temp + + def get_pdf(self): + temp = self.get_tex_dir() + tex_path = os.path.join(temp, 'doc.tex') + try: + cwd = os.getcwd() + except OSError: + cwd = None + os.chdir(temp) + + if self.verbose: + for i in range(self.tex_passes): + p = call(['xelatex', tex_path]) + else: + for i in range(self.tex_passes): + p = call(['xelatex', '-interaction=batchmode', tex_path], + stdout=PIPE, stderr=PIPE) + if p: + raise ParseError("Error parsing .tex file: %s" % tex_path) - return document + if cwd is not None: + os.chdir(cwd) + output_file = NamedTemporaryFile(prefix='librarian', suffix='.pdf', delete=False) + pdf_path = os.path.join(temp, 'doc.pdf') + shutil.move(pdf_path, output_file.name) + shutil.rmtree(temp) + return IOFile.from_filename(output_file.name) -if __name__ == '__main__': - import sys - from librarian import DirDocProvider + def build(self, verbose=False, save_tex=None, morefloats=None): + """ morefloats: new/old/none + """ + self.verbose = verbose + self.save_tex = save_tex - if len(sys.argv) < 2: - print >> sys.stderr, 'Usage: python pdf.py ' - sys.exit(1) + if morefloats is None and package_available('morefloats', 'maxfloats=19'): + morefloats = 'new' + self.morefloats = morefloats - main_input = sys.argv[1] - basepath, ext = os.path.splitext(main_input) - path, slug = os.path.realpath(basepath).rsplit('/', 1) - provider = DirDocProvider(path) - transform(provider, slug, output_dir=path) + book_info = self.wldoc.book_info + if self.cover_class: + self.cover = self.cover_class(book_info) + return self.get_pdf()