X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/d7ba2c607dacf7a6136b83a1588b5adf2278ad46..3cae7a60cf91b6e2b84628c5f3b4b2df77eda5d6:/librarian/pdf.py diff --git a/librarian/pdf.py b/librarian/pdf.py index 1bfd949..d67bddf 100644 --- a/librarian/pdf.py +++ b/librarian/pdf.py @@ -3,17 +3,22 @@ # 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 from copy import deepcopy from subprocess import call, PIPE - -import sys +from itertools import chain from Texml.processor import process from lxml import etree @@ -21,8 +26,10 @@ from lxml.etree import XMLSyntaxError, XSLTApplyError from librarian.dcparser import Person from librarian.parser import WLDocument -from librarian import ParseError, DCNS, get_resource +from librarian import ParseError, DCNS, get_resource, OutputFile, RDFNS from librarian import functions +from librarian.cover import make_cover +from .sponsor import sponsor_logo functions.reg_substitute_entities() @@ -35,19 +42,21 @@ STYLESHEETS = { 'wl2tex': 'pdf/wl2tex.xslt', } -CUSTOMIZATIONS = [ - 'nofootnotes', - 'nothemes', - 'onehalfleading', - 'doubleleading', - 'nowlfont', - ] +# CUSTOMIZATIONS = [ +# 'nofootnotes', +# 'nothemes', +# 'defaultleading', +# 'onehalfleading', +# 'doubleleading', +# 'nowlfont', +# ] + 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'); + >>> t = etree.fromstring('A-B-CX-Y-Z') + >>> insert_tags(t, re.compile('-'), 'd') >>> print etree.tostring(t) ABCXYZ """ @@ -77,7 +86,7 @@ def substitute_hyphens(doc): insert_tags(doc, re.compile("(?<=[^-\s])-(?=[^-\s])"), "dywiz", - exclude=[DCNS("identifier.url"), DCNS("rights.license")] + exclude=[DCNS("identifier.url"), DCNS("rights.license"), "meta"] ) @@ -89,12 +98,38 @@ def fix_hanging(doc): ) +def fix_tables(doc): + for kol in doc.iter(tag='kol'): + if kol.tail is not None: + if not kol.tail.strip(): + kol.tail = None + for table in chain(doc.iter(tag='tabela'), doc.iter(tag='tabelka')): + if table.get('ramka') == '1' or table.get('ramki') == '1': + table.set('_format', '|' + 'X|' * len(table[0])) + else: + table.set('_format', 'X' * len(table[0])) + + +def mark_subauthors(doc): + root_author = ', '.join(elem.text for elem in doc.findall('./' + RDFNS('RDF') + '//' + DCNS('creator_parsed'))) + last_author = None + # jeśli autor jest inny niż autor całości i niż poprzedni autor + # to wstawiamy jakiś znacznik w rdf? + for subutwor in doc.xpath('/utwor/utwor'): + author = ', '.join(elem.text for elem in subutwor.findall('.//' + DCNS('creator_parsed'))) + if author not in (last_author, root_author): + subutwor.find('.//' + RDFNS('RDF')).append(etree.Element('use_subauthor')) + last_author = author + + 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 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'): + 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) @@ -135,10 +170,13 @@ def hack_motifs(doc): def parse_creator(doc): - """ find all dc:creator and dc.contributor tags and add *_parsed versions with forenames first """ - for person in doc.xpath("|".join('//dc:'+(tag) for tag in ( - 'creator', 'contributor.translator', 'contributor.editor', 'contributor.technical_editor')), - namespaces = {'dc': str(DCNS)})[::-1]: + """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) @@ -173,52 +211,59 @@ def package_available(package, args='', verbose=False): return p == 0 -def transform(provider, slug=None, file_path=None, - output_file=None, output_dir=None, make_dir=False, verbose=False, save_tex=None, morefloats=None, - cover=None, flags=None, customizations=None): +def transform(wldoc, verbose=False, save_tex=None, morefloats=None, + cover=None, flags=None, customizations=None, ilustr_path='', latex_dir=False): """ produces a PDF file with XeLaTeX - provider: a DocProvider - slug: slug of file to process, available by provider - file_path can be provided instead of a slug - 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 //.pdf istead of /.pdf + wldoc: a WLDocument verbose: prints all output from LaTeX save_tex: path to save the intermediary LaTeX file to morefloats (old/new/none): force specific morefloats - cover: a cover.Cover object + cover: a cover.Cover factory or True for default flags: less-advertising, customizations: user requested customizations regarding various formatting parameters (passed to wl LaTeX class) """ # Parse XSLT try: - if file_path: - if slug: - raise ValueError('slug or file_path should be specified, not both') - document = load_including_children(provider, file_path=file_path) - else: - if not slug: - raise ValueError('either slug or file_path should be specified') - document = load_including_children(provider, slug=slug) + book_info = wldoc.book_info + document = load_including_children(wldoc) + root = document.edoc.getroot() if cover: - document.edoc.getroot().set('data-cover-width', str(cover.width)) - document.edoc.getroot().set('data-cover-height', str(cover.height)) + if cover is True: + cover = make_cover + bound_cover = cover(book_info, width=1200) + root.set('data-cover-width', str(bound_cover.width)) + root.set('data-cover-height', str(bound_cover.height)) + if bound_cover.uses_dc_cover: + if book_info.cover_by: + root.set('data-cover-by', book_info.cover_by) + if book_info.cover_source: + root.set('data-cover-source', book_info.cover_source) if flags: for flag in flags: - document.edoc.getroot().set('flag-' + flag, 'yes') + root.set('flag-' + flag, 'yes') # check for LaTeX packages if morefloats: - document.edoc.getroot().set('morefloats', morefloats.lower()) + root.set('morefloats', morefloats.lower()) elif package_available('morefloats', 'maxfloats=19'): - document.edoc.getroot().set('morefloats', 'new') + root.set('morefloats', 'new') # add customizations if customizations is not None: - document.edoc.getroot().set('customizations', u','.join(customizations)) + root.set('customizations', u','.join(customizations)) + + # add editors info + editors = document.editors() + if editors: + root.set('editors', u', '.join(sorted( + editor.readable() for editor in editors))) + if document.book_info.funders: + root.set('funders', u', '.join(document.book_info.funders)) + if document.book_info.thanks: + root.set('thanks', document.book_info.thanks) # hack the tree move_motifs_inside(document.edoc) @@ -226,27 +271,39 @@ def transform(provider, slug=None, file_path=None, parse_creator(document.edoc) substitute_hyphens(document.edoc) fix_hanging(document.edoc) - - # find output dir - if make_dir and output_dir is not None: - author = unicode(document.book_info.author) - output_dir = os.path.join(output_dir, author) + fix_tables(document.edoc) + mark_subauthors(document.edoc) # wl -> TeXML style_filename = get_stylesheet("wl2tex") style = etree.parse(style_filename) - - texml = document.transform(style) + functions.reg_mathml_latex() # TeXML -> LaTeX temp = mkdtemp('-wl2pdf') + for ilustr in document.edoc.findall("//ilustr"): + shutil.copy(os.path.join(ilustr_path, ilustr.get("src")), temp) + + for sponsor in book_info.sponsors: + ins = etree.Element("data-sponsor", name=sponsor) + logo = sponsor_logo(sponsor) + if logo: + fname = 'sponsor-%s' % os.path.basename(logo) + shutil.copy(logo, os.path.join(temp, fname)) + ins.set('src', fname) + root.insert(0, ins) + + if book_info.sponsor_note: + root.set("sponsor-note", book_info.sponsor_note) + + texml = document.transform(style) + if cover: - c = cover(document.book_info.author.readable(), document.book_info.title) with open(os.path.join(temp, 'cover.png'), 'w') as f: - c.save(f) + bound_cover.save(f, quality=80) - del document # no longer needed large object :) + del document # no longer needed large object :) tex_path = os.path.join(temp, 'doc.tex') fout = open(tex_path, 'w') @@ -261,68 +318,60 @@ def transform(provider, slug=None, file_path=None, shutil.copy(get_resource('pdf/wl.cls'), temp) shutil.copy(get_resource('res/wl-logo.png'), temp) - cwd = os.getcwd() + if latex_dir: + return temp + + try: + cwd = os.getcwd() + except OSError: + cwd = None os.chdir(temp) - if verbose: - p = call(['xelatex', tex_path]) - else: - p = call(['xelatex', '-interaction=batchmode', tex_path], stdout=PIPE, stderr=PIPE) - if p: - raise ParseError("Error parsing .tex file") + # some things work better when compiled twice + # (table of contents, [line numbers - disabled]) + for run in xrange(2): + if verbose: + p = call(['xelatex', tex_path]) + else: + p = call(['xelatex', '-interaction=batchmode', tex_path], stdout=PIPE, stderr=PIPE) + if p: + raise ParseError("Error parsing .tex file") - os.chdir(cwd) + if cwd is not None: + os.chdir(cwd) - # save the PDF + output_file = NamedTemporaryFile(prefix='librarian', suffix='.pdf', delete=False) pdf_path = os.path.join(temp, 'doc.pdf') - if output_dir is not None: - try: - os.makedirs(output_dir) - except OSError: - pass - if slug: - output_path = os.path.join(output_dir, '%s.pdf' % slug) - else: - output_path = os.path.join(output_dir, os.path.splitext(os.path.basename(file_path))[0] + '.pdf') - shutil.move(pdf_path, output_path) - else: - if hasattr(output_file, 'write'): - # file-like object - with open(pdf_path) as f: - output_file.write(f.read()) - output_file.close() - else: - # path to output file - shutil.copy(pdf_path, output_file) + shutil.move(pdf_path, output_file.name) shutil.rmtree(temp) + return OutputFile.from_filename(output_file.name) except (XMLSyntaxError, XSLTApplyError), e: raise ParseError(e) -def load_including_children(provider, slug=None, uri=None, file_path=None): - """ makes one big xml file with children inserted at end - either slug or uri must be provided +def load_including_children(wldoc=None, provider=None, uri=None): + """ Makes one big xml file with children inserted at end. + + Either wldoc or provider and URI must be provided. """ - if uri: + if uri and provider: f = provider.by_uri(uri) - elif slug: - f = provider[slug] - elif file_path: - f = open(file_path, 'r') + 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 slug, URI nor file path provided for a book.') + raise ValueError('Neither a WLDocument, nor provider and URI were provided.') - text = f.read().decode('utf-8') text = re.sub(ur"([\u0400-\u04ff]+)", ur"\1", text) - document = WLDocument.from_string(text, True, - parse_dublincore=True) + document = WLDocument.from_string(text, parse_dublincore=True, provider=provider) + document.swap_endlines() - f.close() for child_uri in document.book_info.parts: - print child_uri - child = load_including_children(provider, uri=child_uri) + child = load_including_children(provider=provider, uri=child_uri) document.edoc.getroot().append(child.edoc.getroot()) return document