X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/bf257ff992a110c085a50b3aeb5edb4e88697f62..f4cbb15b0bcd2b876872983238a1656fbd990700:/librarian/pdf.py diff --git a/librarian/pdf.py b/librarian/pdf.py index 0485613..5df8dad 100644 --- a/librarian/pdf.py +++ b/librarian/pdf.py @@ -3,10 +3,17 @@ # 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, NamedTemporaryFile import re @@ -21,8 +28,9 @@ from librarian.dcparser import Person from librarian.parser import WLDocument from librarian import ParseError, DCNS, get_resource, OutputFile from librarian import functions -from librarian.cover import ImageCover as WLCover +from librarian.cover import WLCover +import itertools, operator functions.reg_substitute_entities() functions.reg_strip() @@ -77,7 +85,7 @@ 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"), "www"] ) @@ -88,6 +96,29 @@ def fix_hanging(doc): 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 """ @@ -135,9 +166,13 @@ def hack_motifs(doc): def parse_creator(doc): - """ find all dc:creator and dc.contributor tags and add *_parsed versions with forenames first """ + """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', 'contributor.editor', 'contributor.technical_editor')), + 'creator', 'contributor.translator')), namespaces = {'dc': str(DCNS)})[::-1]: if not person.text: continue @@ -173,55 +208,66 @@ def package_available(package, args='', verbose=False): return p == 0 -def transform(wldoc, verbose=False, save_tex=None, morefloats=None, - cover=None, flags=None, customizations=None, - imgdir=""): +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 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 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) + documentclass: LaTeX document class, defaults to wl + resources: a directory with resources, copied to place where LaTeX compilation is made """ # Parse XSLT try: + book_info = wldoc.book_info document = load_including_children(wldoc) + root = document.edoc.getroot() if cover: if cover is True: cover = WLCover - the_cover = cover(document.book_info) - document.edoc.getroot().set('data-cover-width', str(the_cover.width)) - document.edoc.getroot().set('data-cover-height', str(the_cover.height)) - if the_cover.uses_dc_cover: - if document.book_info.cover_by: - document.edoc.getroot().set('data-cover-by', document.book_info.cover_by) - if document.book_info.cover_source: - document.edoc.getroot().set('data-cover-source', document.book_info.cover_source) + 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: - 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)) + + 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) + move_motifs_inside(document.edoc) + hack_motifs(document.edoc) + fake_tables(document.edoc) parse_creator(document.edoc) - if document.book_info.language == 'pol': - substitute_hyphens(document.edoc) + substitute_hyphens(document.edoc) fix_hanging(document.edoc) # wl -> TeXML @@ -229,17 +275,16 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None, style = etree.parse(style_filename) texml = document.transform(style) - etree.dump(texml.getroot()) + + if save_texml: + texml.write(save_texml) + # TeXML -> LaTeX temp = mkdtemp('-wl2pdf') if cover: - with open(os.path.join(temp, 'cover.jpg'), 'w') as f: - the_cover.save(f) - - for img in document.edoc.findall('//ilustr'): - shutil.copy(os.path.join(imgdir, img.get('src')), temp) - + with open(os.path.join(temp, 'cover.png'), 'w') as f: + bound_cover.save(f) del document # no longer needed large object :) @@ -254,20 +299,32 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None, # LaTeX -> PDF 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('logo.eps', temp) - - cwd = os.getcwd() + 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) output_file = NamedTemporaryFile(prefix='librarian', suffix='.pdf', delete=False) pdf_path = os.path.join(temp, 'doc.pdf') @@ -276,7 +333,6 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None, return OutputFile.from_filename(output_file.name) except (XMLSyntaxError, XSLTApplyError), e: - print e raise ParseError(e) @@ -298,7 +354,8 @@ def load_including_children(wldoc=None, provider=None, uri=None): text = re.sub(ur"([\u0400-\u04ff]+)", ur"\1", text) - document = WLDocument.from_string(text, parse_dublincore=True) + document = WLDocument.from_string(text, + parse_dublincore=True, provider=provider) document.swap_endlines() for child_uri in document.book_info.parts: