fixes #2356: set User-Agent

[librarian.git] / librarian / pdf.py

diff --git a/librarian/pdf.py b/librarian/pdf.py
index 0b80a2d..9fb92b1 100644 (file)
--- a/librarian/pdf.py
+++ b/librarian/pdf.py
@@ -3,41 +3,54 @@
 # 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 lxml.etree import XMLSyntaxError, XSLTApplyError
 
 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, OutputFile

 from librarian import functions
 from librarian import functions

-
+from librarian.cover import WLCover

 
 
 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',

 }
 
 }
 

+#CUSTOMIZATIONS = [
+#    'nofootnotes',
+#    'nothemes',
+#    'defaultleading',
+#    'onehalfleading',
+#    'doubleleading',
+#    'nowlfont',
+#    ]

 
 

-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');


@@ -45,46 +58,47 @@ def insert_tags(doc, split_re, tagname):
     <a><b>A<d/>B<d/>C</b>X<d/>Y<d/>Z</a>
     """
 
     <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")]
+                )

 
 
 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")]
+                )

 
 
 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'):
 
 
 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'):

-            print motif.text

             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
             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


@@ -126,20 +140,36 @@ 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.
+    """
+    for person in doc.xpath("|".join('//dc:'+(tag) for tag in (
+                    'creator', 'contributor.translator')),
+                    namespaces = {'dc': str(DCNS)})[::-1]:
+        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}


@@ -153,50 +183,77 @@ 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):
+def transform(wldoc, verbose=False, save_tex=None, morefloats=None,
+              cover=None, flags=None, customizations=None):

     """ produces a PDF file with XeLaTeX
 
     """ produces a PDF file with XeLaTeX
 

-    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
+    wldoc: a WLDocument

     verbose: prints all output from LaTeX
     save_tex: path to save the intermediary LaTeX file to
     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)

     """
 
     # Parse XSLT
     try:
     """
 
     # 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.
-=============================================================================="""
+        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)
 
         # hack the tree
         move_motifs_inside(document.edoc)
         hack_motifs(document.edoc)

+        parse_creator(document.edoc)

         substitute_hyphens(document.edoc)
         fix_hanging(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)
         # wl -> TeXML
         style_filename = get_stylesheet("wl2tex")
         style = etree.parse(style_filename)

+

         texml = document.transform(style)
         texml = document.transform(style)

-        del document # no longer needed large object :)

 
         # TeXML -> LaTeX
         temp = mkdtemp('-wl2pdf')
 
         # TeXML -> LaTeX
         temp = mkdtemp('-wl2pdf')

+
+        if cover:
+            with open(os.path.join(temp, 'cover.png'), 'w') as f:
+                bound_cover.save(f)
+
+        del document # no longer needed large object :)
+

         tex_path = os.path.join(temp, 'doc.tex')
         fout = open(tex_path, 'w')
         process(StringIO(texml), fout, 'utf-8')
         tex_path = os.path.join(temp, 'doc.tex')
         fout = open(tex_path, 'w')
         process(StringIO(texml), fout, 'utf-8')


@@ -207,10 +264,13 @@ Some documents with many motifs in long stanzas or paragraphs may not compile.
             shutil.copy(tex_path, save_tex)
 
         # LaTeX -> PDF
             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)
+        shutil.copy(get_resource('pdf/wl.cls'), temp)
+        shutil.copy(get_resource('res/wl-logo.png'), temp)

 
 

-        cwd = os.getcwd()
+        try:
+            cwd = os.getcwd()
+        except OSError:
+            cwd = None

         os.chdir(temp)
 
         if verbose:
         os.chdir(temp)
 
         if verbose:


@@ -220,50 +280,42 @@ Some documents with many motifs in long stanzas or paragraphs may not compile.
         if p:
             raise ParseError("Error parsing .tex file")
 
         if p:
             raise ParseError("Error parsing .tex file")
 

-        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 OutputFile.from_filename(output_file.name)

 
     except (XMLSyntaxError, XSLTApplyError), e:
         raise ParseError(e)
 
 
 
     except (XMLSyntaxError, XSLTApplyError), e:
         raise ParseError(e)
 
 

-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
+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.

     """
 
     """
 

-    if uri:
+    if uri and provider:

         f = provider.by_uri(uri)
         f = provider.by_uri(uri)

-    elif slug:
-        f = provider[slug]
+        text = f.read().decode('utf-8')
+        f.close()
+    elif wldoc is not None:
+        text = etree.tostring(wldoc.edoc, encoding=unicode)
+        provider = wldoc.provider

     else:
     else:

-        raise ValueError('Neither slug nor URI provided for a book.')
+        raise ValueError('Neither a WLDocument, nor provider and URI were provided.')

 
 

-    document = WLDocument.from_file(f, True,
-        parse_dublincore=True,
-        preserve_lines=False)
+    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:
 
     for child_uri in document.book_info.parts:

-        child = load_including_children(provider, uri=child_uri)
+        child = load_including_children(provider=provider, uri=child_uri)

         document.edoc.getroot().append(child.edoc.getroot())
         document.edoc.getroot().append(child.edoc.getroot())

-

     return document
     return document