a2x.py

#!/usr/bin/env python3

"""
a2x - A toolchain manager for AsciiDoc (converts Asciidoc text files to other
      file formats)

Copyright: Stuart Rackham (c) 2009
License:   MIT
Email:     srackham@gmail.com
"""

import os
import fnmatch
from html.parser import HTMLParser
import re
import shutil
import subprocess
import sys
import traceback
from urllib.parse import urlparse
import zipfile
import xml.dom.minidom
import mimetypes

PROG = os.path.basename(os.path.splitext(__file__)[0])
VERSION = '9.0.0rc1'

# AsciiDoc global configuration file directory.
# NOTE: CONF_DIR is "fixed up" by Makefile -- don't rename or change syntax.
CONF_DIR = '/etc/asciidoc'


######################################################################
# Default configuration file parameters.
######################################################################

# Optional environment variable dictionary passed to
# executing programs. If set to None the existing
# environment is used.
ENV = None

# External executables.
ASCIIDOC = 'asciidoc'
XSLTPROC = 'xsltproc'
DBLATEX = 'dblatex'         # pdf generation.
FOP = 'fop'                 # pdf generation (--fop option).
W3M = 'w3m'                 # primary text file generator.
LYNX = 'lynx'               # alternate text file generator.
XMLLINT = 'xmllint'         # Set to '' to disable.
EPUBCHECK = 'epubcheck'     # Set to '' to disable.
# External executable default options.
ASCIIDOC_OPTS = ''
BACKEND_OPTS = ''
DBLATEX_OPTS = ''
FOP_OPTS = ''
LYNX_OPTS = '-dump'
W3M_OPTS = '-dump -cols 70 -T text/html -no-graph'
XSLTPROC_OPTS = ''

######################################################################
# End of configuration file parameters.
######################################################################


#####################################################################
# Utility functions
#####################################################################

OPTIONS = None  # These functions read verbose and dry_run command options.


def errmsg(msg):
    print('%s: %s\n' % (PROG, msg), file=sys.stderr)


def warning(msg):
    errmsg('WARNING: %s' % msg)


def infomsg(msg):
    print('%s: %s' % (PROG, msg))


def die(msg, exit_code=1):
    errmsg('ERROR: %s' % msg)
    sys.exit(exit_code)


def trace():
    """Print traceback to stderr."""
    errmsg('-'*60)
    traceback.print_exc(file=sys.stderr)
    errmsg('-'*60)


def verbose(msg):
    if OPTIONS.verbose or OPTIONS.dry_run:
        infomsg(msg)


class AttrDict(dict):
    """
    Like a dictionary except values can be accessed as attributes i.e. obj.foo
    can be used in addition to obj['foo'].
    If self._default has been set then it will be returned if a non-existent
    attribute is accessed (instead of raising an AttributeError).
    """
    def __getattr__(self, key):
        try:
            return self[key]
        except KeyError as k:
            if '_default' in self:
                return self['_default']
            else:
                raise AttributeError from k

    def __setattr__(self, key, value):
        self[key] = value

    def __delattr__(self, key):
        try:
            del self[key]
        except KeyError as k:
            raise AttributeError from k

    def __repr__(self):
        return '<AttrDict ' + dict.__repr__(self) + '>'

    def __getstate__(self):
        return dict(self)

    def __setstate__(self, value):
        for k, v in value.items():
            self[k] = v


def isexecutable(file_name):
    return os.path.isfile(file_name) and os.access(file_name, os.X_OK)


def find_executable(file_name):
    '''
    Search for executable file_name in the system PATH.
    Return full path name or None if not found.
    '''
    def _find_executable(file_name):
        if os.path.split(file_name)[0] != '':
            # file_name includes directory so don't search path.
            if not isexecutable(file_name):
                return None
            else:
                return file_name
        for p in os.environ.get('PATH', os.defpath).split(os.pathsep):
            f = os.path.join(p, file_name)
            if isexecutable(f):
                return os.path.realpath(f)
        return None
    if os.name == 'nt' and os.path.splitext(file_name)[1] == '':
        for ext in ('.cmd', '.bat', '.exe'):
            result = _find_executable(file_name + ext)
            if result:
                break
    else:
        result = _find_executable(file_name)
    return result


def write_file(filename, data, mode='w', encoding='utf-8'):
    with open(filename, mode=mode, encoding=encoding) as f:
        f.write(data)


def read_file(filename, mode='r', encoding='utf-8'):
    with open(filename, mode=mode, encoding=encoding) as f:
        return f.read()


def shell_cd(path):
    verbose('chdir %s' % path)
    if not OPTIONS.dry_run:
        os.chdir(path)


def shell_makedirs(path):
    if os.path.isdir(path):
        return
    verbose('creating %s' % path)
    if not OPTIONS.dry_run:
        os.makedirs(path)


def shell_copy(src, dst):
    verbose('copying "%s" to "%s"' % (src, dst))
    if not OPTIONS.dry_run:
        shutil.copy(src, dst)


def shell_rm(path):
    if not os.path.exists(path):
        return
    verbose('deleting %s' % path)
    if not OPTIONS.dry_run:
        os.unlink(path)


def shell_rmtree(path):
    if not os.path.isdir(path):
        return
    verbose('deleting %s' % path)
    if not OPTIONS.dry_run:
        shutil.rmtree(path)


def shell(cmd, raise_error=True):
    '''
    Execute command cmd in shell and return tuple
    (stdoutdata, stderrdata, returncode).
    If raise_error is True then a non-zero return terminates the application.
    '''
    if os.name == 'nt':
        # TODO: this is probably unnecessary, see:
        # http://groups.google.com/group/asciidoc/browse_frm/thread/9442ee0c419f1242
        # Windows doesn't like running scripts directly so explicitly
        # specify python interpreter.
        # Extract first (quoted or unquoted) argument.
        mo = re.match(r'^\s*"\s*(?P<arg0>[^"]+)\s*"', cmd)
        if not mo:
            mo = re.match(r'^\s*(?P<arg0>[^ ]+)', cmd)
        if mo.group('arg0').endswith('.py'):
            cmd = 'python ' + cmd
        # Remove redundant quoting -- this is not just cosmetic,
        # quoting seems to dramatically decrease the allowed command
        # length in Windows XP.
        cmd = re.sub(r'"([^ ]+?)"', r'\1', cmd)
    verbose('executing: %s' % cmd)
    if OPTIONS.dry_run:
        return
    stdout = stderr = subprocess.PIPE
    try:
        popen = subprocess.Popen(cmd, stdout=stdout, stderr=stderr,
                                 shell=True, env=ENV, universal_newlines=True)
    except OSError as e:
        die('failed: %s: %s' % (cmd, e))
    stdoutdata, stderrdata = popen.communicate()
    if OPTIONS.verbose:
        print(stdoutdata)
        print(stderrdata)
    if popen.returncode != 0 and raise_error:
        die('%s returned non-zero exit status %d' % (cmd, popen.returncode))
    return (stdoutdata, stderrdata, popen.returncode)


def find_resources(files, tagname, attrname, filter=None):
    '''
    Search all files and return a list of local URIs from attrname attribute
    values in tagname tags.
    Handles HTML open and XHTML closed tags.
    Non-local URIs are skipped.
    files can be a file name or a list of file names.
    The filter function takes a dictionary of tag attributes and returns True
    if the URI is to be included.
    '''
    class FindResources(HTMLParser):
        # Nested parser class shares locals with enclosing function.
        def handle_startendtag(self, tag, attrs):
            self.handle_starttag(tag, attrs)

        def handle_starttag(self, tag, attrs):
            attrs = dict(attrs)
            if tag == tagname and (filter is None or filter(attrs)):
                # Accept only local URIs.
                uri = urlparse(attrs[attrname])
                if uri[0] in ('', 'file') and not uri[1] and uri[2]:
                    result.append(uri[2])

    if isinstance(files, str):
        files = [files]
    result = []
    for filename in files:
        verbose('finding resources in: %s' % filename)
        if OPTIONS.dry_run:
            continue
        parser = FindResources()
        with open(filename, 'rb') as open_file:
            contents = open_file.read()
        mo = re.search(b'\A<\?xml.* encoding="(.*?)"', contents)
        contents = contents.decode(mo.group(1).decode('utf-8') if mo else 'utf-8')
        parser.feed(contents)
        parser.close()
    result = list(set(result))   # Drop duplicate values.
    result.sort()
    return result


# NOT USED.
def copy_files(files, src_dir, dst_dir):
    '''
    Copy list of relative file names from src_dir to dst_dir.
    '''
    for filename in files:
        filename = os.path.normpath(filename)
        if os.path.isabs(filename):
            continue
        src = os.path.join(src_dir, filename)
        dst = os.path.join(dst_dir, filename)
        if not os.path.exists(dst):
            if not os.path.isfile(src):
                warning('missing file: %s' % src)
                continue
            dstdir = os.path.dirname(dst)
            shell_makedirs(dstdir)
            shell_copy(src, dst)


def find_files(path, pattern):
    '''
    Return list of file names matching pattern in directory path.
    '''
    result = []
    for (p, dirs, files) in os.walk(path):
        for f in files:
            if fnmatch.fnmatch(f, pattern):
                result.append(os.path.normpath(os.path.join(p, f)))
    return result


def exec_xsltproc(xsl_file, xml_file, dst_dir, opts=''):
    cwd = os.getcwd()
    shell_cd(dst_dir)
    try:
        shell('"%s" %s "%s" "%s"' % (XSLTPROC, opts, xsl_file, xml_file))
    finally:
        shell_cd(cwd)


def get_source_options(asciidoc_file):
    '''
    Look for a2x command options in AsciiDoc source file.
    Limitation: options cannot contain double-quote characters.
    '''
    def parse_options():
        # Parse options to result sequence.
        inquotes = False
        opt = ''
        for c in options:
            if c == '"':
                if inquotes:
                    result.append(opt)
                    opt = ''
                    inquotes = False
                else:
                    inquotes = True
            elif c == ' ':
                if inquotes:
                    opt += c
                elif opt:
                    result.append(opt)
                    opt = ''
            else:
                opt += c
        if opt:
            result.append(opt)

    result = []
    if os.path.isfile(asciidoc_file):
        options = ''
        with open(asciidoc_file) as f:
            for line in f:
                mo = re.search(r'^//\s*a2x:', line)
                if mo:
                    options += ' ' + line[mo.end():].strip()
        parse_options()
    return result


#####################################################################
# Application class
#####################################################################

class A2X(AttrDict):
    '''
    a2x options and conversion functions.
    '''

    def execute(self):
        '''
        Process a2x command.
        '''
        self.process_options()
        # Append configuration file options.
        self.asciidoc_opts += ' ' + ASCIIDOC_OPTS
        self.dblatex_opts += ' ' + DBLATEX_OPTS
        self.fop_opts += ' ' + FOP_OPTS
        self.xsltproc_opts += ' ' + XSLTPROC_OPTS
        self.backend_opts += ' ' + BACKEND_OPTS
        # Execute to_* functions.
        if self.backend:
            self.to_backend()
        else:
            self.__getattribute__('to_'+self.format)()
        if not (self.keep_artifacts or self.format == 'docbook' or self.skip_asciidoc):
            shell_rm(self.dst_path('.xml'))

    def load_conf(self):
        '''
        Load a2x configuration file from default locations and --conf-file
        option.
        '''
        global ASCIIDOC
        CONF_FILE = 'a2x.conf'
        a2xdir = os.path.dirname(os.path.realpath(__file__))
        conf_files = []
        # From a2x.py directory.
        conf_files.append(os.path.join(a2xdir, CONF_FILE))
        # If the asciidoc executable and conf files are in the a2x directory
        # then use the local copy of asciidoc and skip the global a2x conf.
        asciidoc = os.path.join(a2xdir, 'asciidoc.py')
        asciidoc_conf = os.path.join(a2xdir, 'asciidoc.conf')
        if os.path.isfile(asciidoc) and os.path.isfile(asciidoc_conf):
            self.asciidoc = asciidoc
        else:
            self.asciidoc = None
            # From global conf directory.
            conf_files.append(os.path.join(CONF_DIR, CONF_FILE))
        # From $HOME directory.
        home_dir = os.environ.get('HOME')
        if home_dir is not None:
            conf_files.append(os.path.join(home_dir, '.asciidoc', CONF_FILE))
        # If asciidoc is not local to a2x then search the PATH.
        if not self.asciidoc:
            self.asciidoc = find_executable(ASCIIDOC)
            if not self.asciidoc:
                die('unable to find asciidoc: %s' % ASCIIDOC)
        # From backend plugin directory.
        if self.backend is not None:
            stdout = shell(self.asciidoc + ' --backend list')[0]
            backends = [(i, os.path.split(i)[1]) for i in stdout.splitlines()]
            backend_dir = [i[0] for i in backends if i[1] == self.backend]
            if len(backend_dir) == 0:
                die('missing %s backend' % self.backend)
            if len(backend_dir) > 1:
                die('more than one %s backend' % self.backend)
            verbose('found %s backend directory: %s' %
                    (self.backend, backend_dir[0]))
            conf_files.append(os.path.join(backend_dir[0], 'a2x-backend.py'))
        # From --conf-file option.
        if self.conf_file is not None:
            if not os.path.isfile(self.conf_file):
                die('missing configuration file: %s' % self.conf_file)
            conf_files.append(self.conf_file)
        # From --xsl-file option.
        if self.xsl_file is not None:
            if not os.path.isfile(self.xsl_file):
                die('missing XSL file: %s' % self.xsl_file)
            self.xsl_file = os.path.abspath(self.xsl_file)
        # Load ordered files.
        for f in conf_files:
            if os.path.isfile(f):
                verbose('loading configuration file: %s' % f)
                exec(open(f).read(), globals())

    def process_options(self):
        '''
        Validate and command options and set defaults.
        '''
        if not os.path.isfile(self.asciidoc_file):
            die('missing ASCIIDOC_FILE: %s' % self.asciidoc_file)
        self.asciidoc_file = os.path.abspath(self.asciidoc_file)
        if os.path.splitext(self.asciidoc_file)[1].lower() == '.xml':
            self.skip_asciidoc = True
        else:
            self.skip_asciidoc = False
        if not self.destination_dir:
            self.destination_dir = os.path.dirname(self.asciidoc_file)
        else:
            if not os.path.isdir(self.destination_dir):
                die('missing --destination-dir: %s' % self.destination_dir)
            self.destination_dir = os.path.abspath(self.destination_dir)
            if self.format not in ('chunked', 'epub', 'htmlhelp', 'xhtml', 'manpage'):
                warning('--destination-dir option is only applicable to HTML and manpage based outputs')
        self.resource_dirs = []
        self.resource_files = []
        if self.resource_manifest:
            if not os.path.isfile(self.resource_manifest):
                die('missing --resource-manifest: %s' % self.resource_manifest)
            with open(self.resource_manifest) as f:
                for r in f:
                    self.resources.append(r.strip())
        for r in self.resources:
            r = os.path.expanduser(r)
            r = os.path.expandvars(r)
            if r.endswith('/') or r.endswith('\\'):
                if os.path.isdir(r):
                    self.resource_dirs.append(r)
                else:
                    die('missing resource directory: %s' % r)
            elif os.path.isdir(r):
                self.resource_dirs.append(r)
            elif r.startswith('.') and '=' in r:
                ext, mimetype = r.split('=')
                mimetypes.add_type(mimetype, ext)
            else:
                self.resource_files.append(r)
        for p in (os.path.dirname(self.asciidoc), CONF_DIR):
            for d in ('images', 'stylesheets'):
                d = os.path.join(p, d)
                if os.path.isdir(d):
                    self.resource_dirs.append(d)
        verbose('resource files: %s' % self.resource_files)
        verbose('resource directories: %s' % self.resource_dirs)
        if not self.doctype and self.format == 'manpage':
            self.doctype = 'manpage'
        if self.doctype:
            self.asciidoc_opts += ' --doctype %s' % self.doctype
        for attr in self.attributes:
            self.asciidoc_opts += ' --attribute "%s"' % attr
#        self.xsltproc_opts += ' --nonet'
        if self.verbose:
            self.asciidoc_opts += ' --verbose'
            self.dblatex_opts += ' -V'
        if self.icons or self.icons_dir:
            params = [
                'callout.graphics 1',
                'navig.graphics 1',
                'admon.textlabel 0',
                'admon.graphics 1',
            ]
            if self.icons_dir:
                params += [
                    'admon.graphics.path "%s/"' % self.icons_dir,
                    'callout.graphics.path "%s/callouts/"' % self.icons_dir,
                    'navig.graphics.path "%s/"' % self.icons_dir,
                ]
        else:
            params = [
                'callout.graphics 0',
                'navig.graphics 0',
                'admon.textlabel 1',
                'admon.graphics 0',
            ]
        if self.stylesheet:
            params += ['html.stylesheet "%s"' % self.stylesheet]
        if self.format == 'htmlhelp':
            params += ['htmlhelp.chm "%s"' % self.basename('.chm'),
                       'htmlhelp.hhp "%s"' % self.basename('.hhp'),
                       'htmlhelp.hhk "%s"' % self.basename('.hhk'),
                       'htmlhelp.hhc "%s"' % self.basename('.hhc')]
        if self.doctype == 'book':
            params += ['toc.section.depth 1']
            # Books are chunked at chapter level.
            params += ['chunk.section.depth 0']
        for o in params:
            if o.split()[0]+' ' not in self.xsltproc_opts:
                self.xsltproc_opts += ' --stringparam ' + o
        if self.fop_opts:
            self.fop = True

    def dst_path(self, ext):
        '''
        Return name of file or directory in the destination directory with
        the same name as the asciidoc source file but with extension ext.
        '''
        return os.path.join(self.destination_dir, self.basename(ext))

    def basename(self, ext):
        '''
        Return the base name of the asciidoc source file but with extension
        ext.
        '''
        return os.path.basename(os.path.splitext(self.asciidoc_file)[0]) + ext

    def asciidoc_conf_file(self, path):
        '''
        Return full path name of file in asciidoc configuration files directory.
        Search first the directory containing the asciidoc executable then
        the global configuration file directory.
        '''
        f = os.path.join(os.path.dirname(self.asciidoc), path)
        if not os.path.isfile(f):
            f = os.path.join(CONF_DIR, path)
            if not os.path.isfile(f):
                die('missing configuration file: %s' % f)
        return os.path.normpath(f)

    def xsl_stylesheet(self, file_name=None):
        '''
        Return full path name of file in asciidoc docbook-xsl configuration
        directory.
        If an XSL file was specified with the --xsl-file option then it is
        returned.
        '''
        if self.xsl_file is not None:
            return self.xsl_file
        if not file_name:
            file_name = self.format + '.xsl'
        return self.asciidoc_conf_file(os.path.join('docbook-xsl', file_name))

    def copy_resources(self, html_files, src_dir, dst_dir, resources=[]):
        '''
        Search html_files for images and CSS resource URIs (html_files can
        be a list of file names or a single file name).
        Copy them from the src_dir to the dst_dir.
        If not found in src_dir then recursively search all specified
        resource directories.
        Optional additional resources files can be passed in the resources
        list.
        '''
        resources = resources[:]
        resources += find_resources(
            html_files,
            'link',
            'href',
            lambda attrs: attrs.get('type') == 'text/css'
        )
        resources += find_resources(html_files, 'img', 'src')
        resources += self.resource_files
        resources = list(set(resources))    # Drop duplicates.
        resources.sort()
        for f in resources:
            if '=' in f:
                src, dst = f.split('=')
                if not dst:
                    dst = src
            else:
                src = dst = f
            src = os.path.normpath(src)
            dst = os.path.normpath(dst)
            if os.path.isabs(dst):
                die('absolute resource file name: %s' % dst)
            if dst.startswith(os.pardir):
                die('resource file outside destination directory: %s' % dst)
            src = os.path.join(src_dir, src)
            dst = os.path.join(dst_dir, dst)
            if not os.path.isfile(src):
                for d in self.resource_dirs:
                    d = os.path.join(src_dir, d)
                    found = find_files(d, os.path.basename(src))
                    if found:
                        src = found[0]
                        break
                else:
                    if not os.path.isfile(dst):
                        die('missing resource: %s' % src)
                    continue
            # Arrive here if resource file has been found.
            if os.path.normpath(src) != os.path.normpath(dst):
                dstdir = os.path.dirname(dst)
                shell_makedirs(dstdir)
                shell_copy(src, dst)

    def to_backend(self):
        '''
        Convert AsciiDoc source file to a backend output file using the global
        'to_<backend name>' function (loaded from backend plugin a2x-backend.py
        file).
        Executes the global function in an A2X class instance context.
        '''
        eval('to_%s(self)' % self.backend)

    def to_docbook(self):
        '''
        Use asciidoc to convert asciidoc_file to DocBook.
        args is a string containing additional asciidoc arguments.
        '''
        docbook_file = self.dst_path('.xml')
        if self.skip_asciidoc:
            if not os.path.isfile(docbook_file):
                die('missing docbook file: %s' % docbook_file)
            return
        shell('"%s" --backend docbook -a "a2x-format=%s" %s --out-file "%s" "%s"' %
             (self.asciidoc, self.format, self.asciidoc_opts, docbook_file, self.asciidoc_file))
        if not self.no_xmllint and XMLLINT:
            shell('"%s" --nonet --noout --valid "%s"' % (XMLLINT, docbook_file))

    def to_xhtml(self):
        self.to_docbook()
        docbook_file = self.dst_path('.xml')
        xhtml_file = self.dst_path('.html')
        opts = '%s --output "%s"' % (self.xsltproc_opts, xhtml_file)
        exec_xsltproc(self.xsl_stylesheet(), docbook_file, self.destination_dir, opts)
        src_dir = os.path.dirname(self.asciidoc_file)
        self.copy_resources(xhtml_file, src_dir, self.destination_dir)

    def to_manpage(self):
        self.to_docbook()
        docbook_file = self.dst_path('.xml')
        opts = self.xsltproc_opts
        exec_xsltproc(self.xsl_stylesheet(), docbook_file, self.destination_dir, opts)

    def to_pdf(self):
        if self.fop:
            self.exec_fop()
        else:
            self.exec_dblatex()

    def exec_fop(self):
        self.to_docbook()
        docbook_file = self.dst_path('.xml')
        xsl = self.xsl_stylesheet('fo.xsl')
        fo = self.dst_path('.fo')
        pdf = self.dst_path('.pdf')
        opts = '%s --output "%s"' % (self.xsltproc_opts, fo)
        exec_xsltproc(xsl, docbook_file, self.destination_dir, opts)
        shell('"%s" %s -fo "%s" -pdf "%s"' % (FOP, self.fop_opts, fo, pdf))
        if not self.keep_artifacts:
            shell_rm(fo)

    def exec_dblatex(self):
        self.to_docbook()
        docbook_file = self.dst_path('.xml')
        xsl = self.asciidoc_conf_file(os.path.join('dblatex','asciidoc-dblatex.xsl'))
        sty = self.asciidoc_conf_file(os.path.join('dblatex','asciidoc-dblatex.sty'))
        shell('"%s" -t %s -p "%s" -s "%s" %s "%s"' %
             (DBLATEX, self.format, xsl, sty, self.dblatex_opts, docbook_file))

    def to_dvi(self):
        self.exec_dblatex()

    def to_ps(self):
        self.exec_dblatex()

    def to_tex(self):
        self.exec_dblatex()

    def to_htmlhelp(self):
        self.to_chunked()

    def to_chunked(self):
        self.to_docbook()
        docbook_file = self.dst_path('.xml')
        opts = self.xsltproc_opts
        xsl_file = self.xsl_stylesheet()
        if self.format == 'chunked':
            dst_dir = self.dst_path('.chunked')
        elif self.format == 'htmlhelp':
            dst_dir = self.dst_path('.htmlhelp')
        if 'base.dir ' not in opts:
            opts += ' --stringparam base.dir "%s/"' % os.path.basename(dst_dir)
        # Create content.
        shell_rmtree(dst_dir)
        shell_makedirs(dst_dir)
        exec_xsltproc(xsl_file, docbook_file, self.destination_dir, opts)
        html_files = find_files(dst_dir, '*.html')
        src_dir = os.path.dirname(self.asciidoc_file)
        self.copy_resources(html_files, src_dir, dst_dir)

    def update_epub_manifest(self, opf_file):
        '''
        Scan the OEBPS directory for any files that have not been registered in
        the OPF manifest then add them to the manifest.
        '''
        opf_dir = os.path.dirname(opf_file)
        resource_files = []
        for (p, dirs, files) in os.walk(os.path.dirname(opf_file)):
            for f in files:
                f = os.path.join(p, f)
                if os.path.isfile(f):
                    assert f.startswith(opf_dir)
                    f = '.' + f[len(opf_dir):]
                    f = os.path.normpath(f)
                    if f not in ['content.opf']:
                        resource_files.append(f)
        opf = xml.dom.minidom.parseString(read_file(opf_file))
        manifest_files = []
        manifest = opf.getElementsByTagName('manifest')[0]
        for el in manifest.getElementsByTagName('item'):
            f = el.getAttribute('href')
            f = os.path.normpath(f)
            manifest_files.append(f)
        count = 0
        for f in resource_files:
            if f not in manifest_files:
                count += 1
                verbose('adding to manifest: %s' % f)
                item = opf.createElement('item')
                item.setAttribute('href', f.replace(os.path.sep, '/'))
                item.setAttribute('id', 'a2x-%d' % count)
                mimetype = mimetypes.guess_type(f)[0]
                if mimetype is None:
                    die('unknown mimetype: %s' % f)
                item.setAttribute('media-type', mimetype)
                manifest.appendChild(item)
        if count > 0:
            write_file(opf_file, opf.toxml())

    def to_epub(self):
        self.to_docbook()
        xsl_file = self.xsl_stylesheet()
        docbook_file = self.dst_path('.xml')
        epub_file = self.dst_path('.epub')
        build_dir = epub_file + '.d'
        shell_rmtree(build_dir)
        shell_makedirs(build_dir)
        # Create content.
        exec_xsltproc(xsl_file, docbook_file, build_dir, self.xsltproc_opts)
        # Copy resources referenced in the OPF and resources referenced by the
        # generated HTML (in theory DocBook XSL should ensure they are
        # identical but this is not always the case).
        src_dir = os.path.dirname(self.asciidoc_file)
        dst_dir = os.path.join(build_dir, 'OEBPS')
        opf_file = os.path.join(dst_dir, 'content.opf')
        opf_resources = find_resources(opf_file, 'item', 'href')
        html_files = find_files(dst_dir, '*.html')
        self.copy_resources(html_files, src_dir, dst_dir, opf_resources)
        # Register any unregistered resources.
        self.update_epub_manifest(opf_file)
        # Build epub archive.
        cwd = os.getcwd()
        shell_cd(build_dir)
        try:
            if not self.dry_run:
                zip_archive = zipfile.ZipFile(epub_file, 'w')
                try:
                    # Create and add uncompressed mimetype file.
                    verbose('archiving: mimetype')
                    write_file('mimetype', 'application/epub+zip')
                    zip_archive.write('mimetype', compress_type=zipfile.ZIP_STORED)
                    # Compress all remaining files.
                    for (p, dirs, files) in os.walk('.'):
                        for f in files:
                            f = os.path.normpath(os.path.join(p,f))
                            if f != 'mimetype':
                                verbose('archiving: %s' % f)
                                zip_archive.write(f, compress_type=zipfile.ZIP_DEFLATED)
                finally:
                    zip_archive.close()
            verbose('created archive: %s' % epub_file)
        finally:
            shell_cd(cwd)
        if not self.keep_artifacts:
            shell_rmtree(build_dir)
        if self.epubcheck and EPUBCHECK:
            if not find_executable(EPUBCHECK):
                warning('epubcheck skipped: unable to find executable: %s' % EPUBCHECK)
            else:
                shell('"%s" "%s"' % (EPUBCHECK, epub_file))

    def to_text(self):
        text_file = self.dst_path('.text')
        html_file = self.dst_path('.text.html')
        if self.lynx:
            shell('"%s" %s --conf-file "%s" -b html4 -a "a2x-format=%s" -o "%s" "%s"' %
                 (self.asciidoc, self.asciidoc_opts, self.asciidoc_conf_file('text.conf'),
                  self.format, html_file, self.asciidoc_file))
            cmd = '"%s" %s "%s" > "%s"' % (LYNX, LYNX_OPTS, html_file, text_file)
            shell(cmd)
        else:
            # Use w3m(1).
            self.to_docbook()
            docbook_file = self.dst_path('.xml')
            opts = '%s --output "%s"' % (self.xsltproc_opts, html_file)
            exec_xsltproc(self.xsl_stylesheet(), docbook_file,
                    self.destination_dir, opts)
            cmd = '"%s" %s "%s" > "%s"' % (W3M, W3M_OPTS, html_file, text_file)
            shell(cmd)
        if not self.keep_artifacts:
            shell_rm(html_file)


#####################################################################
# Script main line.
#####################################################################

if __name__ == '__main__':
    description = '''A toolchain manager for AsciiDoc (converts Asciidoc text files to other file formats)'''
    from argparse import ArgumentParser
    parser = ArgumentParser(usage='usage: %(prog)s [OPTIONS] ASCIIDOC_FILE',
        description=description)
    parser.add_argument('--version', action='version', version='{} {}'.format(PROG, VERSION))
    parser.add_argument('-a', '--attribute',
        action='append', dest='attributes', default=[], metavar='ATTRIBUTE',
        help='set asciidoc attribute value')
    parser.add_argument('--asciidoc-opts',
        action='append', dest='asciidoc_opts', default=[],
        metavar='ASCIIDOC_OPTS', help='asciidoc options')
    #DEPRECATED
    parser.add_argument('--copy',
        action='store_true', dest='copy', default=False,
        help='DEPRECATED: does nothing')
    parser.add_argument('--conf-file',
        dest='conf_file', default=None, metavar='CONF_FILE',
        help='configuration file')
    parser.add_argument('-D', '--destination-dir',
        action='store', dest='destination_dir', default=None, metavar='PATH',
        help='output directory (defaults to ASCIIDOC_FILE directory)')
    parser.add_argument('-d','--doctype',
        action='store', dest='doctype', metavar='DOCTYPE',
        choices=('article','manpage','book'),
        help='article, manpage, book')
    parser.add_argument('-b','--backend',
        action='store', dest='backend', metavar='BACKEND',
        help='name of backend plugin')
    parser.add_argument('--epubcheck',
        action='store_true', dest='epubcheck', default=False,
        help='check EPUB output with epubcheck')
    parser.add_argument('-f','--format',
        action='store', dest='format', metavar='FORMAT', default = 'pdf',
        choices=('chunked','epub','htmlhelp','manpage','pdf', 'text',
                 'xhtml','dvi','ps','tex','docbook'),
        help='chunked, epub, htmlhelp, manpage, pdf, text, xhtml, dvi, ps, tex, docbook')
    parser.add_argument('--icons',
        action='store_true', dest='icons', default=False,
        help='use admonition, callout and navigation icons')
    parser.add_argument('--icons-dir',
        action='store', dest='icons_dir',
        default=None, metavar='PATH',
        help='admonition and navigation icon directory')
    parser.add_argument('-k', '--keep-artifacts',
        action='store_true', dest='keep_artifacts', default=False,
        help='do not delete temporary build files')
    parser.add_argument('--lynx',
        action='store_true', dest='lynx', default=False,
        help='use lynx to generate text files')
    parser.add_argument('-L', '--no-xmllint',
        action='store_true', dest='no_xmllint', default=False,
        help='do not check asciidoc output with xmllint')
    parser.add_argument('-n','--dry-run',
        action='store_true', dest='dry_run', default=False,
        help='just print the commands that would have been executed')
    parser.add_argument('-r','--resource',
        action='append', dest='resources', default=[],
        metavar='PATH',
        help='resource file or directory containing resource files')
    parser.add_argument('-m', '--resource-manifest',
        action='store', dest='resource_manifest', default=None, metavar='FILE',
        help='read resources from FILE')
    #DEPRECATED
    parser.add_argument('--resource-dir',
        action='append', dest='resources', default=[],
        metavar='PATH',
        help='DEPRECATED: use --resource')
    #DEPRECATED
    parser.add_argument('-s','--skip-asciidoc',
        action='store_true', dest='skip_asciidoc', default=False,
        help='DEPRECATED: redundant')
    parser.add_argument('--stylesheet',
        action='store', dest='stylesheet', default=None,
        metavar='STYLESHEET',
        help='HTML CSS stylesheet file name')
    #DEPRECATED
    parser.add_argument('--safe',
        action='store_true', dest='safe', default=False,
        help='DEPRECATED: does nothing')
    parser.add_argument('--dblatex-opts',
        action='append', dest='dblatex_opts', default=[],
        metavar='DBLATEX_OPTS', help='dblatex options')
    parser.add_argument('--backend-opts',
        action='append', dest='backend_opts', default=[],
        metavar='BACKEND_OPTS', help='backend plugin options')
    parser.add_argument('--fop',
        action='store_true', dest='fop', default=False,
        help='use FOP to generate PDF files')
    parser.add_argument('--fop-opts',
        action='append', dest='fop_opts', default=[],
        metavar='FOP_OPTS', help='options for FOP pdf generation')
    parser.add_argument('--xsltproc-opts',
        action='append', dest='xsltproc_opts', default=[],
        metavar='XSLTPROC_OPTS', help='xsltproc options for XSL stylesheets')
    parser.add_argument('--xsl-file',
        action='store', dest='xsl_file', metavar='XSL_FILE',
        help='custom XSL stylesheet')
    parser.add_argument('-v', '--verbose',
        action='count', dest='verbose', default=0,
        help='increase verbosity')
    parser.add_argument("asciidoc_file", action="store", help="AsciiDoc source file")
    source_options = get_source_options(sys.argv[-1])
    argv = source_options + sys.argv[1:]
    opts = parser.parse_args(argv)
    opts.asciidoc_opts = ' '.join(opts.asciidoc_opts)
    opts.dblatex_opts = ' '.join(opts.dblatex_opts)
    opts.fop_opts = ' '.join(opts.fop_opts)
    opts.xsltproc_opts = ' '.join(opts.xsltproc_opts)
    opts.backend_opts = ' '.join(opts.backend_opts)
    opts = vars(opts)  # Convert argparse.Values to dict.
    a2x = A2X(opts)
    OPTIONS = a2x           # verbose and dry_run used by utility functions.
    verbose('args: %r' % argv)
    a2x.asciidoc_file = opts['asciidoc_file']
    try:
        a2x.load_conf()
        a2x.execute()
    except KeyboardInterrupt:
        sys.exit(1)