X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/1ce2c1255aee01fab9940fc26d251767bbf8c960..f4cbb15b0bcd2b876872983238a1656fbd990700:/librarian/pdf.py?ds=inline diff --git a/librarian/pdf.py b/librarian/pdf.py index 07272f2..5df8dad 100644 --- a/librarian/pdf.py +++ b/librarian/pdf.py @@ -3,27 +3,34 @@ # 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 distutils.dir_util import copy_tree # shutil.copytree is so uncapable. 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 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, DCNS +from librarian import ParseError, DCNS, get_resource, OutputFile from librarian import functions +from librarian.cover import WLCover +import itertools, operator functions.reg_substitute_entities() functions.reg_strip() @@ -32,12 +39,20 @@ functions.reg_ends_white() functions.reg_texcommand() STYLESHEETS = { - 'wl2tex': 'xslt/wl2tex.xslt', + 'wl2tex': 'pdf/wl2tex.xslt', } +#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 + """ 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'); @@ -67,20 +82,43 @@ def insert_tags(doc, split_re, tagname, exclude=None): def substitute_hyphens(doc): - insert_tags(doc, + insert_tags(doc, re.compile("(?<=[^-\s])-(?=[^-\s])"), "dywiz", - exclude=[DCNS("identifier.url"), DCNS("rights.license")] + exclude=[DCNS("identifier.url"), DCNS("rights.license"), "www"] ) def fix_hanging(doc): - insert_tags(doc, + insert_tags(doc, re.compile("(?<=\s\w)\s+"), "nbsp", exclude=[DCNS("identifier.url"), DCNS("rights.license")] ) +def fake_tables(doc): + for tabela in doc.findall("//tabela"): + # are we dealing with a table of proper structure? + # two levels of same tags, and all tags on second level + # must be of same count. + def tag_count(m, k): + m[k.tag] = m.get(k.tag, 0) + 1 + return m + + child_tags = reduce(tag_count, list(tabela), {}) + if len(child_tags) != 1: + return + grandchild_tags = reduce(tag_count, itertools.chain(*[list(c) for c in tabela]), {}) + if len(grandchild_tags) != 1: + return + if len(set(grandchild_tags.values())) != 1: + return + + for row in tabela: + row.tag = 'r' + for col in row: + col.tag = 'c' + return def move_motifs_inside(doc): """ moves motifs to be into block elements """ @@ -128,29 +166,35 @@ def hack_motifs(doc): def parse_creator(doc): - """ find all dc:creator tags and add dc:creator_parsed with forenames first """ - for creator in doc.findall('//'+DCNS('creator')): - p = Person.from_text(creator.text) - creator_parsed = deepcopy(creator) - creator_parsed.tag = DCNS('creator_parsed') - creator_parsed.text = ' '.join(p.first_names + (p.last_name,)) - creator.getparent().insert(0, creator_parsed) + """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 package_available(package, args='', verbose=False): - """ check if a verion of a latex package accepting given args is available """ + """ 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{book} + \documentclass{wl} \usepackage[%s]{%s} \begin{document} \end{document} @@ -164,58 +208,86 @@ 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): +def transform(wldoc, verbose=False, save_tex=None, save_texml=None, morefloats=None, + cover=None, cover_file=None, flags=None, customizations=None, documentclass='wl', resources=None, tworuns=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 + save_texml: path to save the intermediary TeXML file to morefloats (old/new/none): force specific morefloats + 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) + documentclass: LaTeX document class, defaults to wl + resources: a directory with resources, copied to place where LaTeX compilation is made """ # 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: + if cover is True: + cover = WLCover + bound_cover = cover(book_info) + 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: + 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: + root.set('customizations', u','.join(customizations)) + + root.set('documentclass', documentclass or 'wl') + + # add editors info + 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) + fake_tables(document.edoc) 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) - # wl -> TeXML style_filename = get_stylesheet("wl2tex") style = etree.parse(style_filename) + texml = document.transform(style) - del document # no longer needed large object :) + + if save_texml: + texml.write(save_texml) # TeXML -> LaTeX temp = mkdtemp('-wl2pdf') + + if cover: + with open(os.path.join(temp, 'cover.png'), 'w') as f: + bound_cover.save(f) + + del document # no longer needed large object :) + tex_path = os.path.join(temp, 'doc.tex') fout = open(tex_path, 'w') process(StringIO(texml), fout, 'utf-8') @@ -226,72 +298,67 @@ def transform(provider, slug=None, file_path=None, shutil.copy(tex_path, save_tex) # LaTeX -> PDF - shutil.copy(get_resource('pdf/wl.sty'), temp) - shutil.copy(get_resource('pdf/wl-logo.png'), temp) - - cwd = os.getcwd() + shutil.copy(get_resource('pdf/wl.cls'), temp) + shutil.copy(get_resource('pdf/wlpub.cls'), temp) + shutil.copy(get_resource('pdf/fnprep.cls'), temp) + shutil.copy(get_resource('res/wl-logo.png'), temp) + shutil.copy(get_resource('res/cover.jpg'), temp) + if resources: + copy_tree(resources, 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") + if resources: + os.putenv("TEXINPUTS", "::.:%s" % resources) + + for run in range(1 + (tworuns and 1 or 0)): + 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) - - f.close() + document = WLDocument.from_string(text, + parse_dublincore=True, provider=provider) + document.swap_endlines() for child_uri in document.book_info.parts: - 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