X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/8059679abd4a0a2e554457dd6266941a31356f73..5175c4cb9704f442c20abffbd575fb588278b3b2:/librarian/pdf.py diff --git a/librarian/pdf.py b/librarian/pdf.py index 9fb92b1..c02e98c 100644 --- a/librarian/pdf.py +++ b/librarian/pdf.py @@ -21,13 +21,11 @@ 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, DCNS, get_resource, OutputFile +from librarian import ParseError, DCNS, get_resource, IOFile, Format from librarian import functions -from librarian.cover import WLCover functions.reg_substitute_entities() @@ -40,20 +38,12 @@ STYLESHEETS = { '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 - >>> 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 """ @@ -83,7 +73,7 @@ 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'] ) @@ -91,16 +81,20 @@ def fix_hanging(doc): insert_tags(doc, re.compile("(?<=\s\w)\s+"), "nbsp", - exclude=[DCNS("identifier.url"), DCNS("rights.license")] + exclude=[DCNS("identifier.url"), DCNS("rights.license"), DCNS("title"), DCNS("description")] ) 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'): + main_tags = ('powiesc', 'opowiadanie', 'liryka_l', 'liryka_lp', + 'dramat_wierszowany_l', 'dramat_wierszowany_lp', 'dramat_wspolczesny') + for master in doc.xpath('|'.join('//' + tag for tag in main_tags)): 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'): + special_tags = ('sekcja_swiatlo', 'sekcja_asterysk', 'separator_linia', + 'begin', 'end', 'motyw', 'extra', 'uwaga') + if sib.tag not in special_tags: # motif shouldn't have a tail - it would be untagged text motif.tail = None motif.getparent().remove(motif) @@ -146,9 +140,10 @@ def parse_creator(doc): 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]: + persons = doc.xpath( + "|".join('//dc:' + tag for tag in ('creator', 'contributor.translator')), + namespaces={'dc': str(DCNS)})[::-1] + for person in persons: if not person.text: continue p = Person.from_text(person.text) @@ -183,102 +178,104 @@ 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): - """ produces a PDF file with XeLaTeX +# not used +def load_including_children(wldoc=None, provider=None, uri=None): + """ Makes one big xml file with children inserted at end. - 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 factory or True for default - flags: less-advertising, - customizations: user requested customizations regarding various formatting parameters (passed to wl LaTeX class) + Either wldoc or provider and URI must be provided. """ - # 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 - 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: - root.set('morefloats', morefloats.lower()) - elif package_available('morefloats', 'maxfloats=19'): - root.set('morefloats', 'new') - - # add customizations - if customizations is not None: - root.set('customizations', u','.join(customizations)) - - # 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) - parse_creator(document.edoc) - substitute_hyphens(document.edoc) - fix_hanging(document.edoc) - - # wl -> TeXML - style_filename = get_stylesheet("wl2tex") - style = etree.parse(style_filename) - - texml = document.transform(style) - - # TeXML -> LaTeX - temp = mkdtemp('-wl2pdf') + if uri and provider: + f = provider.by_uri(uri) + # WTF DocProvider.by_uri() returns IOFile, so no .read() there + 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.') - if cover: - with open(os.path.join(temp, 'cover.png'), 'w') as f: - bound_cover.save(f) + text = re.sub(ur"([\u0400-\u04ff]+)", ur"\1", text) - del document # no longer needed large object :) + document = WLDocument.from_string(text, parse_dublincore=True, provider=provider) + document.swap_endlines() - tex_path = os.path.join(temp, 'doc.tex') - fout = open(tex_path, 'w') - process(StringIO(texml), fout, 'utf-8') - fout.close() - del texml + 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 + + +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. + + """ + + 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 - if save_tex: - shutil.copy(tex_path, save_tex) + @property + def customization_str(self): + """ For use in XSLT. """ + return u','.join(k for k, v in self.customization.items() if v) - # LaTeX -> PDF + def get_texml(self): + raise NotImplementedError + + 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(get_resource('res/wl-logo.png'), 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 verbose: - p = call(['xelatex', tex_path]) + p = None + if self.verbose: + for i in xrange(self.tex_passes): + p = call(['xelatex', tex_path]) else: - p = call(['xelatex', '-interaction=batchmode', tex_path], stdout=PIPE, stderr=PIPE) + for i in xrange(self.tex_passes): + p = call(['xelatex', '-interaction=batchmode', tex_path], stdout=PIPE, stderr=PIPE) if p: - raise ParseError("Error parsing .tex file") + raise ParseError("Error parsing .tex file: %s" % tex_path) if cwd is not None: os.chdir(cwd) @@ -287,35 +284,20 @@ def transform(wldoc, verbose=False, save_tex=None, morefloats=None, pdf_path = os.path.join(temp, 'doc.pdf') shutil.move(pdf_path, output_file.name) shutil.rmtree(temp) - return OutputFile.from_filename(output_file.name) + return IOFile.from_filename(output_file.name) - except (XMLSyntaxError, XSLTApplyError), e: - raise ParseError(e) + def build(self, verbose=False, save_tex=None, morefloats=None): + """ morefloats: new/old/none + """ + self.verbose = verbose + self.save_tex = save_tex + if morefloats is None and package_available('morefloats', 'maxfloats=19'): + morefloats = 'new' + self.morefloats = morefloats -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. - """ + book_info = self.wldoc.book_info + if self.cover_class: + self.cover = self.cover_class(book_info) - 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.') - - text = re.sub(ur"([\u0400-\u04ff]+)", ur"\1", text) - - 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=provider, uri=child_uri) - document.edoc.getroot().append(child.edoc.getroot()) - return document + return self.get_pdf()