X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/ee1f5f9be7d4e2bb843c01a118cc090ab97e6cc6..fc2ee5e934d0da3535b9be659bbf67ebe9437b47:/librarian/epub.py diff --git a/librarian/epub.py b/librarian/epub.py index a0d9742..ef07c3f 100644 --- a/librarian/epub.py +++ b/librarian/epub.py @@ -7,41 +7,21 @@ from __future__ import with_statement import os import os.path -import shutil -import sys +import subprocess from copy import deepcopy from lxml import etree import zipfile +from tempfile import mkdtemp +from shutil import rmtree -from librarian import XMLNamespace, RDFNS, DCNS, WLNS, XHTMLNS, NoDublinCore -from librarian.parser import WLDocument - -#TODO: shouldn't be repeated here -NCXNS = XMLNamespace("http://www.daisy.org/z3986/2005/ncx/") -OPFNS = XMLNamespace("http://www.idpf.org/2007/opf") - - -class DocProvider(object): - class DoesNotExist(Exception): - pass - - def by_slug(self, slug): - raise NotImplemented - - def __getitem__(self, slug): - return self.by_slug(slug) - - def by_uri(self, uri): - return self.by_slug(uri.rsplit('/', 1)[1]) +import sys +from librarian import XMLNamespace, RDFNS, DCNS, WLNS, NCXNS, OPFNS, NoDublinCore +from librarian.dcparser import BookInfo -class DirDocProvider(DocProvider): - def __init__(self, dir): - self.dir = dir - self.files = {} +from librarian import functions - def by_slug(self, slug): - return open(os.path.join(self.dir, '%s.xml' % slug)) +functions.reg_person_name() def inner_xml(node): @@ -63,7 +43,6 @@ def set_inner_xml(node, text): xyz """ - p = etree.fromstring('%s' % text) node.text = p.text node[:] = p[:] @@ -103,12 +82,11 @@ def replace_characters(node): def replace_chars(text): if text is None: return None - return text.replace("&", "&")\ - .replace("---", "—")\ - .replace("--", "–")\ - .replace(",,", "„")\ - .replace('"', "”")\ - .replace("'", "’") + return text.replace("---", u"\u2014")\ + .replace("--", u"\u2013")\ + .replace(",,", u"\u201E")\ + .replace('"', u"\u201D")\ + .replace("'", u"\u2019") if node.tag == 'extra': node.clear() else: @@ -234,6 +212,14 @@ class TOC(object): return counter +def used_chars(element): + """ Lists characters used in an ETree Element """ + chars = set((element.text or '') + (element.tail or '')) + for child in element: + chars = chars.union(used_chars(child)) + return chars + + def chop(main_text): """ divide main content of the XML file into chunks """ @@ -259,7 +245,7 @@ def chop(main_text): def transform_chunk(chunk_xml, chunk_no, annotations): - """ transforms one chunk, returns a HTML string and a TOC object """ + """ transforms one chunk, returns a HTML string, a TOC object and a set of used characters """ toc = TOC() for element in chunk_xml[0]: @@ -270,15 +256,20 @@ def transform_chunk(chunk_xml, chunk_no, annotations): element.set('sub', str(subnumber)) find_annotations(annotations, chunk_xml, chunk_no) replace_by_verse(chunk_xml) - output_html = etree.tostring(xslt(chunk_xml, res('xsltScheme.xsl')), pretty_print=True) - return output_html, toc + html_tree = xslt(chunk_xml, res('xsltScheme.xsl')) + chars = used_chars(html_tree.getroot()) + output_html = etree.tostring(html_tree, pretty_print=True) + return output_html, toc, chars -def transform(provider, slug, output_file): - """ produces an epub +def transform(provider, slug=None, file_path=None, output_file=None, output_dir=None, make_dir=False, verbose=False): + """ produces a EPUB file - provider is a DocProvider - output_file should be filelike object + provider: a DocProvider + slug: slug of file to process, available by provider + 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 //.epub instead of /.epub """ def transform_file(input_xml, chunk_counter=1, first=True): @@ -289,14 +280,19 @@ def transform(provider, slug, output_file): # every input file will have a TOC entry, # pointing to starting chunk toc = TOC(node_name(input_xml.find('.//'+DCNS('title'))), chunk_counter) + chars = set() if first: # write book title page + html_tree = xslt(input_xml, res('xsltTitle.xsl')) + chars = used_chars(html_tree.getroot()) zip.writestr('OPS/title.html', - etree.tostring(xslt(input_xml, res('xsltTitle.xsl')), pretty_print=True)) + etree.tostring(html_tree, pretty_print=True)) elif children: # write title page for every parent + html_tree = xslt(input_xml, res('xsltChunkTitle.xsl')) + chars = used_chars(html_tree.getroot()) zip.writestr('OPS/part%d.html' % chunk_counter, - etree.tostring(xslt(input_xml, res('xsltChunkTitle.xsl')), pretty_print=True)) + etree.tostring(html_tree, pretty_print=True)) add_to_manifest(manifest, chunk_counter) add_to_spine(spine, chunk_counter) chunk_counter += 1 @@ -313,9 +309,10 @@ def transform(provider, slug, output_file): if main_text is not None: replace_characters(main_text) - for chunk_no, chunk_xml in enumerate(chop(main_text), chunk_counter): - chunk_html, chunk_toc = transform_chunk(chunk_xml, chunk_counter, annotations) + for chunk_xml in chop(main_text): + chunk_html, chunk_toc, chunk_chars = transform_chunk(chunk_xml, chunk_counter, annotations) toc.extend(chunk_toc) + chars = chars.union(chunk_chars) zip.writestr('OPS/part%d.html' % chunk_counter, chunk_html) add_to_manifest(manifest, chunk_counter) add_to_spine(spine, chunk_counter) @@ -324,11 +321,43 @@ def transform(provider, slug, output_file): if children: for child in children: child_xml = etree.parse(provider.by_uri(child)) - child_toc, chunk_counter = transform_file(child_xml, chunk_counter, first=False) + child_toc, chunk_counter, chunk_chars = transform_file(child_xml, chunk_counter, first=False) toc.append(child_toc) + chars = chars.union(chunk_chars) - return toc, chunk_counter + return toc, chunk_counter, chars + # read metadata from the first file + if file_path: + if slug: + raise ValueError('slug or file_path should be specified, not both') + f = open(file_path, 'r') + input_xml = etree.parse(f) + f.close() + else: + if not slug: + raise ValueError('either slug or file_path should be specified') + input_xml = etree.parse(provider[slug]) + + metadata = input_xml.find('.//'+RDFNS('Description')) + if metadata is None: + raise NoDublinCore('Document has no DublinCore - which is required.') + book_info = BookInfo.from_element(input_xml) + metadata = etree.ElementTree(metadata) + + # if output to dir, create the file + if output_dir is not None: + if make_dir: + author = unicode(book_info.author) + output_dir = os.path.join(output_dir, author) + try: + os.makedirs(output_dir) + except OSError: + pass + if slug: + output_file = open(os.path.join(output_dir, '%s.epub' % slug), 'w') + else: + output_file = open(os.path.join(output_dir, os.path.splitext(os.path.basename(file_path))[0] + '.epub'), 'w') zip = zipfile.ZipFile(output_file, 'w', zipfile.ZIP_DEFLATED) @@ -346,12 +375,6 @@ def transform(provider, slug, output_file): for fname in 'style.css', 'logo_wolnelektury.png': zip.write(res(fname), os.path.join('OPS', fname)) - # metadata from first file - input_xml = etree.parse(provider[slug]) - metadata = input_xml.find('.//'+RDFNS('Description')) - if metadata is None: - raise NoDublinCore('Document has no DublinCore - which is required.') - metadata = etree.ElementTree(metadata) opf = xslt(metadata, res('xsltContent.xsl')) manifest = opf.find('.//' + OPFNS('manifest')) spine = opf.find('.//' + OPFNS('spine')) @@ -364,13 +387,14 @@ def transform(provider, slug, output_file): 'version="2005-1">' \ '' \ 'Strona tytułowa' \ - '' \ - 'Początek utworu' \ '') nav_map = toc_file[-1] - toc, chunk_counter = transform_file(input_xml) - toc_counter = toc.write_to_xml(nav_map, 3) # we already have 2 navpoints + toc, chunk_counter, chars = transform_file(input_xml) + + if not toc.children: + toc.add(u"Początek utworu", 1) + toc_counter = toc.write_to_xml(nav_map, 2) # Last modifications in container files and EPUB creation if len(annotations) > 0: @@ -382,8 +406,26 @@ def transform(provider, slug, output_file): spine.append(etree.fromstring( '')) replace_by_verse(annotations) + html_tree = xslt(annotations, res("xsltAnnotations.xsl")) + chars = chars.union(used_chars(html_tree.getroot())) zip.writestr('OPS/annotations.html', etree.tostring( - xslt(annotations, res("xsltAnnotations.xsl")), pretty_print=True)) + html_tree, pretty_print=True)) + + # strip fonts + tmpdir = mkdtemp('-librarian-epub') + cwd = os.getcwd() + + os.chdir(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'font-optimizer')) + for fname in 'DejaVuSerif.ttf', 'DejaVuSerif-Bold.ttf', 'DejaVuSerif-Italic.ttf', 'DejaVuSerif-BoldItalic.ttf': + optimizer_call = ['perl', 'subset.pl', '--chars', ''.join(chars).encode('utf-8'), res('../fonts/' + fname), os.path.join(tmpdir, fname)] + if verbose: + print "Running font-optimizer" + subprocess.check_call(optimizer_call) + else: + subprocess.check_call(optimizer_call, stdout=subprocess.PIPE, stderr=subprocess.PIPE) + zip.write(os.path.join(tmpdir, fname), os.path.join('OPS', fname)) + rmtree(tmpdir) + os.chdir(cwd) zip.writestr('OPS/content.opf', etree.tostring(opf, pretty_print=True)) contents = [] @@ -399,17 +441,3 @@ def transform(provider, slug, output_file): set_inner_xml(toc_file[1], ''.join(('', title, ''))) zip.writestr('OPS/toc.ncx', etree.tostring(toc_file, pretty_print=True)) zip.close() - - -if __name__ == '__main__': - if len(sys.argv) < 2: - print >> sys.stderr, 'Usage: python epub.py ' - sys.exit(1) - - main_input = sys.argv[1] - basepath, ext = os.path.splitext(main_input) - path, slug = os.path.realpath(basepath).rsplit('/', 1) - output = basepath + '.epub' - provider = DirDocProvider(path) - transform(provider, slug, open(output, 'w')) -