X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/a67b6e3ad2751083a433228454665c212c9a5a1e..1e438727014e352799b4dfb8d3dfb2f8e7a58251:/librarian/pdf.py

diff --git a/librarian/pdf.py b/librarian/pdf.py
index b9ead15..1989239 100644
--- a/librarian/pdf.py
+++ b/librarian/pdf.py
@@ -11,32 +11,32 @@ from StringIO import StringIO
 from tempfile import mkdtemp
 import re
 from copy import deepcopy
+from subprocess import call, PIPE
 
 import sys
-sys.path.append('..') # for running from working copy
 
 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
+from librarian import ParseError, DCNS, get_resource
 from librarian import functions
 
 
-
 functions.reg_substitute_entities()
-functions.reg_person_name()
 functions.reg_strip()
 functions.reg_starts_white()
 functions.reg_ends_white()
+functions.reg_texcommand()
 
 STYLESHEETS = {
     'wl2tex': 'xslt/wl2tex.xslt',
 }
 
 
-def insert_tags(doc, split_re, tagname):
+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>');
@@ -45,127 +45,227 @@ def insert_tags(doc, split_re, tagname):
     <a><b>A<d/>B<d/>C</b>X<d/>Y<d/>Z</a>
     """
 
-    for elem in doc.iter():
+    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)
-            elem.text = chunks.pop(0)
-            while chunks:
+            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
-            elem.tail = chunks.pop(0)
-            while chunks:
+            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):
     insert_tags(doc, 
                 re.compile("(?<=[^-\s])-(?=[^-\s])"),
-                "dywiz")
+                "dywiz",
+                exclude=[DCNS("identifier.url"), DCNS("rights.license")]
+                )
 
 
 def fix_hanging(doc):
     insert_tags(doc, 
                 re.compile("(?<=\s\w)\s+"),
-                "nbsp")
+                "nbsp",
+                exclude=[DCNS("identifier.url"), DCNS("rights.license")]
+                )
+
+
+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'):
+        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'):
+                    # motif shouldn't have a tail - it would be untagged text
+                    motif.tail = None
+                    motif.getparent().remove(motif)
+                    sib.insert(0, motif)
+                    break
+
+
+def hack_motifs(doc):
+    """ dirty hack for the marginpar-creates-orphans LaTeX problem
+    see http://www.latex-project.org/cgi-bin/ltxbugs2html?pr=latex/2304
 
+    moves motifs in stanzas from first verse to second
+    and from next to last to last, then inserts negative vspace before them
+    """
+    for motif in doc.findall('//strofa//motyw'):
+        # find relevant verse-level tag
+        verse, stanza = motif, motif.getparent()
+        while stanza is not None and stanza.tag != 'strofa':
+            verse, stanza = stanza, stanza.getparent()
+        breaks_before = sum(1 for i in verse.itersiblings('br', preceding=True))
+        breaks_after = sum(1 for i in verse.itersiblings('br'))
+        if (breaks_before == 0 and breaks_after > 0) or breaks_after == 1:
+            move_by = 1
+            if breaks_after == 2:
+                move_by += 1
+            moved_motif = deepcopy(motif)
+            motif.tag = 'span'
+            motif.text = None
+            moved_motif.tail = None
+            moved_motif.set('moved', str(move_by))
+
+            for br in verse.itersiblings('br'):
+                if move_by > 1:
+                    move_by -= 1
+                    continue
+                br.addnext(moved_motif)
+                break
+
+
+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)
 
-def get_resource(path):
-    return os.path.join(os.path.dirname(__file__), path)
 
 def get_stylesheet(name):
     return get_resource(STYLESHEETS[name])
 
-def transform(provider, slug, output_file=None, output_dir=None):
-    """ produces a pdf file
 
-    provider is a DocProvider
-    either output_file (a file-like object) or output_dir (path to file/dir) should be specified
-    if output_dir is specified, file will be written to <output_dir>/<author>/<slug>.pdf
+def package_available(package, args='', verbose=False):
+    """ 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}
+        \usepackage[%s]{%s}
+        \begin{document}
+        \end{document}
+        """ % (args, package))
+    f.close()
+    if verbose:
+        p = call(['xelatex', '-output-directory', tempdir, fpath])
+    else:
+        p = call(['xelatex', '-interaction=batchmode', '-output-directory', tempdir, fpath], stdout=PIPE, stderr=PIPE)
+    shutil.rmtree(tempdir)
+    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):
+    """ 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 <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
+    morefloats (old/new/none): force specific morefloats
     """
 
     # Parse XSLT
     try:
-        style_filename = get_stylesheet("wl2tex")
-        style = etree.parse(style_filename)
-
-        document = load_including_children(provider, slug)
-
-        # dirty hack for the marginpar-creates-orphans LaTeX problem
-        # see http://www.latex-project.org/cgi-bin/ltxbugs2html?pr=latex/2304
-        for motif in document.edoc.findall('//strofa//motyw'):
-            # find relevant verse-level tag
-            verse, stanza = motif, motif.getparent()
-            while stanza is not None and stanza.tag != 'strofa':
-                verse, stanza = stanza, stanza.getparent()
-            breaks_before = sum(1 for i in verse.itersiblings('br', preceding=True))
-            breaks_after = sum(1 for i in verse.itersiblings('br'))
-            if (breaks_before == 0 and breaks_after > 0) or breaks_after == 1:
-                move_by = 1
-                if breaks_after == 2:
-                    move_by += 1
-                moved_motif = deepcopy(motif)
-                motif.tag = 'span'
-                motif.text = None
-                moved_motif.tail = None
-                moved_motif.set('moved', str(move_by))
-
-                for br in verse.itersiblings(tag='br'):
-                    if move_by > 1:
-                        move_by -= 1
-                        continue
-                    br.addnext(moved_motif)
-                    break
-
+        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)
+
+        # check for LaTeX packages
+        if morefloats:
+            document.edoc.getroot().set('morefloats', morefloats.lower())
+        elif package_available('morefloats', 'maxfloats=19'):
+            document.edoc.getroot().set('morefloats', 'new')
+
+        # hack the tree
+        move_motifs_inside(document.edoc)
+        hack_motifs(document.edoc)
+        parse_creator(document.edoc)
         substitute_hyphens(document.edoc)
         fix_hanging(document.edoc)
 
-        # if output to dir, create the file
-        if output_dir is not None:
+        # 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 :)
 
-        temp = mkdtemp('wl2pdf-')
+        # TeXML -> LaTeX
+        temp = mkdtemp('-wl2pdf')
         tex_path = os.path.join(temp, 'doc.tex')
         fout = open(tex_path, 'w')
-        process(StringIO(texml), fout, 'utf8', 255, 0, 0)
+        process(StringIO(texml), fout, 'utf-8')
         fout.close()
         del texml
 
+        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)
-        print "pdflatex -output-directory %s %s" % (temp, os.path.join(temp, 'doc.tex'))
-        if os.system("pdflatex -output-directory %s %s" % (temp, os.path.join(temp, 'doc.tex'))):
+
+        cwd = os.getcwd()
+        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")
 
+        os.chdir(cwd)
+
+        # save the 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)
+            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:
-            with open(pdf_path) as f:
-                output_file.write(f.read())
-            output_file.close()
+            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.rmtree(temp)
 
-        return True
     except (XMLSyntaxError, XSLTApplyError), e:
         raise ParseError(e)
 
 
-def load_including_children(provider, slug=None, uri=None):
+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
     """
@@ -174,31 +274,21 @@ def load_including_children(provider, slug=None, uri=None):
         f = provider.by_uri(uri)
     elif slug:
         f = provider[slug]
+    elif file_path:
+        f = open(file_path, 'r')
     else:
-        raise ValueError('Neither slug nor URI provided for a book.')
+        raise ValueError('Neither slug, URI nor file path provided for a book.')
 
-    document = WLDocument.from_file(f, True,
-        parse_dublincore=True,
-        preserve_lines=False)
+    text = f.read().decode('utf-8')
+    text = re.sub(ur"([\u0400-\u04ff]+)", ur"<alien>\1</alien>", text)
+
+    document = WLDocument.from_string(text, True,
+        parse_dublincore=True)
+
+    f.close()
 
     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
-
-
-if __name__ == '__main__':
-    import sys
-    from librarian import DirDocProvider
-
-    if len(sys.argv) < 2:
-        print >> sys.stderr, 'Usage: python pdf.py <input file>'
-        sys.exit(1)
-
-    main_input = sys.argv[1]
-    basepath, ext = os.path.splitext(main_input)
-    path, slug = os.path.realpath(basepath).rsplit('/', 1)
-    provider = DirDocProvider(path)
-    transform(provider, slug, output_dir=path)
-