Refactor collection of text and comments nodes

[extfs.d.git] / xml-minidom

#! /usr/bin/env python
"""XML Virtual FileSystem for Midnight Commander

The script requires Midnight Commander 3.1+
(http://www.midnight-commander.org/), Python 2.4+ (http://www.python.org/).

For mc 4.7+ put the script in $HOME/[.local/share/].mc/extfs.d.
For older versions put it in /usr/[local/][lib|share]/mc/extfs
and add a line "xml" to the /usr/[local/][lib|share]/mc/extfs/extfs.ini.
Make the script executable.

For mc 4.7+ run this "cd" command in the Midnight Commander (in the "bindings"
file the command is "%cd"): cd file/xml://; In older versions it is
cd file#xml, where "file" is the name of your XML file.

The VFS represents tags as directories; the directories are numbered to
distinguish tags with the same name; also numbering helps to sort tags by their
order in XML instead of sorting them by name. Attributes, text nodes and
comments are represented as text files; attributes are shown in a file named
"attributes", attributes are listed in the file as name=value lines (I
deliberately ignore a small chance of newline characters in values); names and
values are reencoded to the console encoding. Text nodes and comments are
collected in a file named "text", stripped and reencoded. The filesystem is
read-only.

It is useful to have a top-down view on an XML structure but it's especially
convenient to extract text values from tags. One can get, for example, a
base64-encoded image - just walk down the VFS to the tag's directory and copy
its text file to a real file.

The VFS was inspired by a FUSE xmlfs: https://github.com/halhen/xmlfs

"""

__version__ = "0.3.2"
__author__ = "Oleg Broytman <phd@phdru.name>"
__copyright__ = "Copyright (C) 2013 PhiloSoft Design"
__license__ = "GPL"

import math
import sys
import xml.dom.minidom

try:
   import locale
   use_locale = True
except ImportError:
   use_locale = False

if use_locale:
   # Get the default charset.
   try:
      lcAll = locale.getdefaultlocale()
   except locale.Error, err:
      print >>sys.stderr, "WARNING:", err
      lcAll = []

   if len(lcAll) == 2:
      default_encoding = lcAll[1]
   else:
      try:
         default_encoding = locale.getpreferredencoding()
      except locale.Error, err:
         print >>sys.stderr, "WARNING:", err
         default_encoding = sys.getdefaultencoding()
else:
   default_encoding = sys.getdefaultencoding()

import logging
logger = logging.getLogger('xml-mcextfs')
log_err_handler = logging.StreamHandler(sys.stderr)
logger.addHandler(log_err_handler)
logger.setLevel(logging.INFO)

if len(sys.argv) < 3:
    logger.critical("""\
XML Virtual FileSystem for Midnight Commander version %s
Author: %s
%s

This is not a program. Put the script in $HOME/[.local/share/].mc/extfs.d or
/usr/[local/][lib|share]/mc/extfs. For more information read the source!""",
   __version__, __author__, __copyright__
)
    sys.exit(1)


locale.setlocale(locale.LC_ALL, '')

def _attrs2text(attrs):
    attrs = [attrs.item(i) for i in range (attrs.length)]
    return '\n'.join(["%s=%s" %
        (a.name.encode(default_encoding, "replace"),
        a.value.encode(default_encoding, "replace"))
        for a in attrs])

def _collect_text(node):
    text_accumulator = []
    for element in node.childNodes:
        if element.localName:
            continue
        elif element.nodeType == element.COMMENT_NODE:
            text = u"<!--%s-->" % element.nodeValue
        elif element.nodeType == element.TEXT_NODE:
            text = element.nodeValue.strip()
        else:
            xml_error("Unknown node type %d" % element.nodeType)
        if text: text_accumulator.append(text)
    return '\n'.join(text_accumulator).encode(default_encoding, "replace")

def _list(node, path=''):
    childNodes = node.childNodes
    n = 0
    for element in childNodes:
        if element.localName:
            n += 1
    if n:
        width = int(math.log10(n))+1
        template = "%%0%dd" % width
    else:
        template = "%d"
    n = 0
    for element in childNodes:
        if element.localName:
            n += 1
            if path:
                subpath = '%s/%s %s' % (path, template % n, element.localName)
            else:
                subpath = '%s %s' % (template % n, element.localName)
            subpath_encoded = subpath.encode(default_encoding, "replace")
            print "dr-xr-xr-x 1 user group 0 Jan 1 00:00 %s" % subpath_encoded
            attrs = element.attributes
            if attrs:
                attr_text = _attrs2text(attrs)
                print "-r--r--r-- 1 user group %d Jan 1 00:00 %s/attributes" % (
                    len(attr_text), subpath_encoded)
            text = _collect_text(element)
            if text:
                print "-r--r--r-- 1 user group %d Jan 1 00:00 %s/text" % (
                    len(text), subpath_encoded)
            _list(element, subpath)

def mcxml_list():
    """List the entire VFS"""

    dom = xml.dom.minidom.parse(sys.argv[2])
    _list(dom)


def _get_child_node(node, i):
    n = 0
    for element in node.childNodes:
        if element.localName:
            n += 1
            if n == i:
                return element
    xml_error('There are less than %d nodes' % i)

def mcxml_copyout():
    """Extract a file from the VFS"""

    node = xml.dom.minidom.parse(sys.argv[2])
    xml_filename = sys.argv[3]
    real_filename = sys.argv[4]

    for path_comp in xml_filename.split('/'):
        if ' ' in path_comp:
            i = int(path_comp.split(' ', 1)[0])
            node = _get_child_node(node, i)
        elif path_comp in ('attributes', 'text'):
            break
        else:
            xml_error('Unknown file')

    if path_comp == 'attributes':
        attrs = node.attributes
        if attrs:
            text = _attrs2text(attrs)
        else:
            xml_error('There are no attributes')

    if path_comp == 'text':
        text = _collect_text(node)

    outfile = open(real_filename, 'w')
    outfile.write(text)
    outfile.close()


def mcxml_copyin():
    """Put a file to the VFS"""
    sys.exit("XML VFS doesn't support adding files (read-only filesystem)")

def mcxml_rm():
    """Remove a file from the VFS"""
    sys.exit("XML VFS doesn't support removing files/directories (read-only filesystem)")

mcxml_rmdir = mcxml_rm

def mcxml_mkdir():
    """Create a directory in the VFS"""
    sys.exit("XML VFS doesn't support creating directories (read-only filesystem)")


def xml_error(error_str):
    logger.critical("Error walking XML file: %s", error_str)
    sys.exit(1)

command = sys.argv[1]
procname = "mcxml_" + command

g = globals()
if not g.has_key(procname):
    logger.critical("Unknown command %s", command)
    sys.exit(1)

try:
    g[procname]()
except SystemExit:
    raise
except:
    logger.exception("Error during run")