accept missing dc tags

[librarian.git] / librarian / pdf.py

diff --git a/librarian/pdf.py b/librarian/pdf.py
index 0b80a2d..c02e98c 100644 (file)
--- a/librarian/pdf.py
+++ b/librarian/pdf.py
@@ -3,90 +3,98 @@
 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
 # 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 __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 re
 from copy import deepcopy
 from subprocess import call, PIPE
 

-import sys
-

 from Texml.processor import process
 from lxml import etree
 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.parser import WLDocument

-from librarian import ParseError
+from librarian import ParseError, DCNS, get_resource, IOFile, Format

 from librarian import functions
 
 
 from librarian import functions
 
 

-

 functions.reg_substitute_entities()
 functions.reg_substitute_entities()

-functions.reg_person_name()

 functions.reg_strip()
 functions.reg_starts_white()
 functions.reg_ends_white()
 functions.reg_strip()
 functions.reg_starts_white()
 functions.reg_ends_white()

+functions.reg_texcommand()

 
 STYLESHEETS = {
 
 STYLESHEETS = {

-    'wl2tex': 'xslt/wl2tex.xslt',
+    'wl2tex': 'pdf/wl2tex.xslt',

 }
 
 
 }
 
 

-def insert_tags(doc, split_re, tagname):
-    """ inserts <tagname> for every occurence of `split_re' in text nodes in the `doc' tree 
+def insert_tags(doc, split_re, tagname, exclude=None):
+    """ inserts <tagname> for every occurence of `split_re' in text nodes in the `doc' tree

 
 

-    >>> t = etree.fromstring('<a><b>A-B-C</b>X-Y-Z</a>');
-    >>> insert_tags(t, re.compile('-'), 'd');
+    >>> t = etree.fromstring('<a><b>A-B-C</b>X-Y-Z</a>')
+    >>> insert_tags(t, re.compile('-'), 'd')

     >>> print etree.tostring(t)
     <a><b>A<d/>B<d/>C</b>X<d/>Y<d/>Z</a>
     """
 
     >>> print etree.tostring(t)
     <a><b>A<d/>B<d/>C</b>X<d/>Y<d/>Z</a>
     """
 

-    for elem in doc.iter():
-        try:
-            if elem.text:
-                chunks = split_re.split(elem.text)
-                while len(chunks) > 1:
-                    ins = etree.Element(tagname)
-                    ins.tail = chunks.pop()
-                    elem.insert(0, ins)
-                elem.text = chunks.pop(0)
-            if elem.tail:
-                chunks = split_re.split(elem.tail)
-                parent = elem.getparent()
-                ins_index = parent.index(elem) + 1
-                while len(chunks) > 1:
-                    ins = etree.Element(tagname)
-                    ins.tail = chunks.pop()
-                    parent.insert(ins_index, ins)
-                elem.tail = chunks.pop(0)
-        except TypeError, e:
-            # element with no children, like comment
-            pass
+    for elem in doc.iter(tag=etree.Element):
+        if exclude and elem.tag in exclude:
+            continue
+        if elem.text:
+            chunks = split_re.split(elem.text)
+            while len(chunks) > 1:
+                ins = etree.Element(tagname)
+                ins.tail = chunks.pop()
+                elem.insert(0, ins)
+            elem.text = chunks.pop(0)
+        if elem.tail:
+            chunks = split_re.split(elem.tail)
+            parent = elem.getparent()
+            ins_index = parent.index(elem) + 1
+            while len(chunks) > 1:
+                ins = etree.Element(tagname)
+                ins.tail = chunks.pop()
+                parent.insert(ins_index, ins)
+            elem.tail = chunks.pop(0)

 
 
 def substitute_hyphens(doc):
 
 
 def substitute_hyphens(doc):

-    insert_tags(doc, 
+    insert_tags(doc,

                 re.compile("(?<=[^-\s])-(?=[^-\s])"),
                 re.compile("(?<=[^-\s])-(?=[^-\s])"),

-                "dywiz")
+                "dywiz",
+                exclude=[DCNS("identifier.url"), DCNS("rights.license"), 'www']
+                )

 
 
 def fix_hanging(doc):
 
 
 def fix_hanging(doc):

-    insert_tags(doc, 
+    insert_tags(doc,

                 re.compile("(?<=\s\w)\s+"),
                 re.compile("(?<=\s\w)\s+"),

-                "nbsp")
+                "nbsp",
+                exclude=[DCNS("identifier.url"), DCNS("rights.license"), DCNS("title"), DCNS("description")]
+                )

 
 
 def move_motifs_inside(doc):
     """ moves motifs to be into block elements """
 
 
 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 motif in master.xpath('motyw'):

-            print motif.text

             for sib in motif.itersiblings():
             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)
                     # motif shouldn't have a tail - it would be untagged text
                     motif.tail = None
                     motif.getparent().remove(motif)


@@ -126,20 +134,37 @@ def hack_motifs(doc):
                 break
 
 
                 break
 
 

-def get_resource(path):
-    return os.path.join(os.path.dirname(__file__), path)
+def parse_creator(doc):
+    """Generates readable versions of creator and translator tags.
+
+    Finds all dc:creator and dc.contributor.translator tags
+    and adds *_parsed versions with forenames first.
+    """
+    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)
+        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_stylesheet(name):
     return get_resource(STYLESHEETS[name])
 
 
 def package_available(package, args='', verbose=False):
 
 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"""
     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}
         \usepackage[%s]{%s}
         \begin{document}
         \end{document}


@@ -153,117 +178,126 @@ def package_available(package, args='', verbose=False):
     return p == 0
 
 
     return p == 0
 
 

-def transform(provider, slug, output_file=None, output_dir=None, make_dir=False, verbose=False, save_tex=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.

 
 

-    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 <output_dir>/<author>/<slug>.pdf istead of <output_dir>/<slug>.pdf
-    verbose: prints all output from LaTeX
-    save_tex: path to save the intermediary LaTeX file to
+    Either wldoc or provider and URI must be provided.

     """
 
     """
 

-    # Parse XSLT
-    try:
-        document = load_including_children(provider, slug)
-
-        # check for latex packages
-        if not package_available('morefloats', 'maxfloats=19', verbose=verbose):
-            document.edoc.getroot().set('old-morefloats', 'yes')
-            print >> sys.stderr, """
-==============================================================================
-LaTeX `morefloats' package is older than v.1.0c or not available at all.
-Some documents with many motifs in long stanzas or paragraphs may not compile.
-=============================================================================="""
-
-        # hack the tree
-        move_motifs_inside(document.edoc)
-        hack_motifs(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 :)
-
-        # TeXML -> LaTeX
-        temp = mkdtemp('-wl2pdf')
-        tex_path = os.path.join(temp, 'doc.tex')
-        fout = open(tex_path, 'w')
-        process(StringIO(texml), fout, 'utf-8')
-        fout.close()
-        del texml
+    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.')
+
+    text = re.sub(ur"([\u0400-\u04ff]+)", ur"<alien>\1</alien>", 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

 
 

-        if save_tex:
-            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)
+class PDFFormat(Format):
+    """ Base PDF format.

 
 

-        cwd = os.getcwd()
+    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
+
+    @property
+    def customization_str(self):
+        """ For use in XSLT. """
+        return u','.join(k for k, v in self.customization.items() if v)
+
+    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(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)
 
         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:
         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:
         if p:

-            raise ParseError("Error parsing .tex file")
+            raise ParseError("Error parsing .tex file: %s" % tex_path)

 
 

-        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')
         pdf_path = os.path.join(temp, 'doc.pdf')

-        if output_dir is not None:
-            try:
-                os.makedirs(output_dir)
-            except OSError:
-                pass
-            output_path = os.path.join(output_dir, '%s.pdf' % slug)
-            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)
         shutil.rmtree(temp)

+        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(provider, slug=None, uri=None):
-    """ makes one big xml file with children inserted at end 
-    either slug or uri must be provided
-    """
-
-    if uri:
-        f = provider.by_uri(uri)
-    elif slug:
-        f = provider[slug]
-    else:
-        raise ValueError('Neither slug nor URI provided for a book.')
-
-    document = WLDocument.from_file(f, True,
-        parse_dublincore=True,
-        preserve_lines=False)
+        book_info = self.wldoc.book_info
+        if self.cover_class:
+            self.cover = self.cover_class(book_info)

 
 

-    for child_uri in document.book_info.parts:
-        child = load_including_children(provider, uri=child_uri)
-        document.edoc.getroot().append(child.edoc.getroot())
-
-    return document
+        return self.get_pdf()