""" path.py - An object representing a path to a file or directory.

Example:

from path import path
d = path('/home/guido/bin')
for f in d.files('*.py'):
    f.chmod(0755)

This module requires Python 2.2 or later.


URL:     http://www.jorendorff.com/articles/python/path
Author:  Jason Orendorff <jason@jorendorff.com> (and others - see the url!)
Date:    23 Feb 2003
"""


# TODO
#   - Is __iter__ worth the trouble?  It breaks the sequence
#     protocol and breaks compatibility with str/unicode.
#   - Perhaps support arguments to touch().
#   - Note:  __add__() technically has a bug, I think, where
#     it doesn't play nice with other types that implement
#     __radd__().  Test this.
#   - Better error message in listdir() when self isn't a
#     directory. (On Windows, the error message really sucks.)
#   - Make sure everything has a good docstring.

from __future__ import generators

import sys, os, fnmatch, glob, shutil, codecs

__version__ = '1.2'
__all__ = ['path']

# Pre-2.3 support.  Are unicode filenames supported?
_base = str
try:
    if os.path.supports_unicode_filenames:
        _base = unicode
except AttributeError:
    pass

# Pre-2.3 workaround for basestring.
try:
    basestring
except NameError:
    basestring = (str, unicode)

# Universal newline support
_textmode = 'r'
if hasattr(file, 'newlines'):
    _textmode = 'U'


class path(_base):
    """ Represents a filesystem path.

    For documentation on individual methods, consult their
    counterparts in os.path.
    """

    # --- Special Python methods.

    def __repr__(self):
        return 'path(%s)' % _base.__repr__(self)

    def __iter__(self):
        return iter(self.listdir())

    # Adding a path and a string yields a path.
    def __add__(self, more):
        return path(_base(self) + more)

    def __radd__(self, other):
        return path(other + _base(self))

    # The / operator joins paths.
    def __div__(self, rel):
        """ fp.__div__(rel) == fp / rel == fp.joinpath(rel)

        Join two path components, adding a separator character if
        needed.
        """
        return path(os.path.join(self, rel))

    # Make the / operator work even when true division is enabled.
    __truediv__ = __div__

    def getcwd():
        """ Return the current working directory as a path object. """
        return path(os.getcwd())
    getcwd = staticmethod(getcwd)


    # --- Operations on path strings.

    def abspath(self):       return path(os.path.abspath(self))
    def normcase(self):      return path(os.path.normcase(self))
    def normpath(self):      return path(os.path.normpath(self))
    def realpath(self):      return path(os.path.realpath(self))
    def expanduser(self):    return path(os.path.expanduser(self))
    def expandvars(self):    return path(os.path.expandvars(self))
    def dirname(self):       return path(os.path.dirname(self))
    basename = os.path.basename

    def expand(self):
        """ Clean up a filename by calling expandvars(),
        expanduser(), and normpath() on it.

        This is commonly everything needed to clean up a filename
        read from a configuration file, for example.
        """
        return self.expandvars().expanduser().normpath()


    def _get_ext(self):
        f, ext = os.path.splitext(_base(self))
        return ext

    def _get_drive(self):
        drive, r = os.path.splitdrive(self)
        return path(drive)

    parent = property(dirname)
    name = property(basename)
    ext = property(
        _get_ext, None, None,
        """ The file extension, for example '.py'. """)
    drive = property(
        _get_drive, None, None,
        """ The drive specifier, for example 'C:'.
        This is always empty on systems that don't use drive specifiers. """)

    def splitpath(self):
        """ p.splitpath() -> Return (p.parent, p.name). """
        parent, child = os.path.split(self)
        return path(parent), child

    def splitdrive(self):
        drive, rel = os.path.splitdrive(self)
        return path(drive), rel

    def splitext(self):
        # Cast to plain string using _base because Python 2.2
        # implementations of os.path.splitext use "for c in path:..."
        # which means something different when applied to a path
        # object.
        filename, ext = os.path.splitext(_base(self))
        return path(filename), ext

    if hasattr(os.path, 'splitunc'):
        def splitunc(self):
            unc, rest = os.path.splitunc(self)
            return path(unc), rest

        def _get_uncshare(self):
            unc, r = os.path.splitunc(self)
            return path(unc)

        uncshare = property(
            _get_uncshare, None, None,
            """ The UNC mount point for this path.
            This is empty for paths on local drives. """)

    def joinpath(self, *args):
        """ Join two or more path components, adding a separator
        character (os.sep) if needed.  Returns a new path
        object.
        """
        return path(os.path.join(self, *args))

    def splitall(self):
        """ Return a list of the path components in this path.

        The first item in the list will be a path.  Its value will be
        either os.curdir, os.pardir, empty, or the root directory of
        this path (for example, '/' or 'C:\\').  The other items in
        the list will be strings.

        path.path.joinpath(*result) will yield the original path.
        """
        parts = []
        loc = self
        while loc != os.curdir and loc != os.pardir:
            prev = loc
            loc, child = prev.splitpath()
            if loc == prev:
                break
            parts.append(child)
        parts.append(loc)
        parts.reverse()
        return parts

    def relpath(self):
        """ Return this path as a relative path,
        based from the current working directory.
        """
        cwd = path(os.getcwd())
        return cwd.relpathto(self)

    def relpathto(self, dest):
        """ Return a relative path from self to dest.

        If there is no relative path from self to dest, for example if
        they reside on different drives in Windows, then this returns
        dest.abspath().
        """
        origin = self.abspath()
        dest = path(dest).abspath()

        orig_list = origin.normcase().splitall()
        # Don't normcase dest!  We want to preserve the case.
        dest_list = dest.splitall()

        if orig_list[0] != os.path.normcase(dest_list[0]):
            # Can't get here from there.
            return dest

        # Find the location where the two paths start to differ.
        i = 0
        for start_seg, dest_seg in zip(orig_list, dest_list):
            if start_seg != os.path.normcase(dest_seg):
                break
            i += 1

        # Now i is the point where the two paths diverge.
        # Need a certain number of "os.pardir"s to work up
        # from the origin to the point of divergence.
        segments = [os.pardir] * (len(orig_list) - i)
        # Need to add the diverging part of dest_list.
        segments += dest_list[i:]
        if len(segments) == 0:
            # If they happen to be identical, use os.curdir.
            return path(os.curdir)
        else:
            return path(os.path.join(*segments))


    # --- Listing, searching, walking, and matching

    def listdir(self, pattern=None):
        """ D.listdir() -> List of items in this directory.

        Use D.files() or D.dirs() instead if you want a listing
        of just files or just subdirectories.

        The elements of the list are path objects.

        With the optional 'pattern' argument, this only lists
        items whose names match the given pattern.
        """
        names = os.listdir(self)
        if pattern is not None:
            names = fnmatch.filter(names, pattern)
        return [self / child for child in names]

    def dirs(self, pattern=None):
        """ D.dirs() -> List of this directory's subdirectories.

        The elements of the list are path objects.
        This does not walk recursively into subdirectories
        (but see path.walkdirs).

        With the optional 'pattern' argument, this only lists
        directories whose names match the given pattern.  For
        example, d.dirs('build-*').
        """
        return [p for p in self.listdir(pattern) if p.isdir()]

    def files(self, pattern=None):
        """ D.files() -> List of the files in this directory.

        The elements of the list are path objects.
        This does not walk into subdirectories (see path.walkfiles).

        With the optional 'pattern' argument, this only lists files
        whose names match the given pattern.  For example,
        d.files('*.pyc').
        """
        
        return [p for p in self.listdir(pattern) if p.isfile()]

    def walk(self, pattern=None):
        """ D.walk() -> iterator over files and subdirs, recursively.

        The iterator yields path objects naming each child item of
        this directory and its descendants.  This requires that
        D.isdir().

        This performs a depth-first traversal of the directory tree.
        Each directory is returned just before all its children.
        """
        for child in self:
            if pattern is None or child.fnmatch(pattern):
                yield child
            if child.isdir():
                for item in child.walk(pattern):
                    yield item

    def walkdirs(self, pattern=None):
        """ D.walkdirs() -> iterator over subdirs, recursively.

        With the optional 'pattern' argument, this yields only
        directories whose names match the given pattern.  For
        example, mydir.walkdirs('*test') yields only directories
        with names ending in 'test'.
        """
        for child in self:
            if child.isdir():
                if pattern is None or child.fnmatch(pattern):
                    yield child
                for subsubdir in child.walkdirs(pattern):
                    yield subsubdir

    def walkfiles(self, pattern=None):
        """ D.walkfiles() -> iterator over files in D, recursively.

        The optional argument, pattern, limits the results to files
        with names that match the pattern.  For example,
        mydir.walkfiles('*.tmp') yields only files with the .tmp
        extension.
        """
        for child in self:
            if child.isfile():
                if pattern is None or child.fnmatch(pattern):
                    yield child
            elif child.isdir():
                for f in child.walkfiles(pattern):
                    yield f

    def fnmatch(self, pattern):
        """ Return True if self.name matches the given pattern.

        pattern - A filename pattern with wildcards,
            for example '*.py'.
        """
        return fnmatch.fnmatch(self.name, pattern)

    def glob(self, pattern):
        """ Return a list of path objects that match the pattern.

        pattern - a path relative to this directory, with wildcards.

        For example, path('/users').glob('*/bin/*') returns a list
        of all the files users have in their bin directories.
        """
        return map(path, glob.glob(_base(self / pattern)))


    # --- Reading an entire file at once.

    def bytes(self):
        """ Open this file, read all bytes, return them as a string. """
        f = file(self, 'rb')
        try:
            return f.read()
        finally:
            f.close()

    def text(self, encoding=None, errors='strict'):
        """ Open this file, read it in, return the content as a string.

        This uses 'U' mode in Python 2.3 and later, so '\r\n' and '\r'
        are automatically translated to '\n'.

        Optional arguments:

        encoding - The Unicode encoding (or character set) of
            the file.  If present, the content of the file is
            decoded and returned as a unicode object; otherwise
            it is returned as an 8-bit str.
        errors - How to handle Unicode errors; see help(str.decode)
            for the options.  Default is 'strict'.
        """
        if encoding is None:
            # 8-bit
            f = file(self, _textmode)
            try:
                return f.read()
            finally:
                f.close()
        else:
            # Unicode
            f = codecs.open(self, 'r', encoding, errors)
            # (Note - Can't use 'U' mode here, since codecs.open
            # doesn't support 'U' mode, even in Python 2.3.)
            try:
                t = f.read()
            finally:
                f.close()
            return t.replace(u'\r\n', u'\n').replace(u'\r', u'\n')

    def lines(self, encoding=None, errors='strict', retain=True):
        """ Open this file, read all lines, return them in a list.

        Optional arguments:
            encoding - The Unicode encoding (or character set) of
                the file.  The default is None, meaning the content
                of the file is read as 8-bit characters and returned
                as a list of (non-Unicode) str objects.
            errors - How to handle Unicode errors; see help(str.decode)
                for the options.  Default is 'strict'
            retain - If true, retain newline characters; but all newline
                character combinations ('\r', '\n', '\r\n') are
                translated to '\n'.  If false, newline characters are
                stripped off.  Default is True.

        This uses 'U' mode in Python 2.3 and later.
        """
        if encoding is None and retain:
            f = file(self, _textmode)
            try:
                return f.readlines()
            finally:
                f.close()
        else:
            return self.text(encoding, errors).splitlines(retain)


    # --- Methods for querying the filesystem.

    exists = os.path.exists
    isabs = os.path.isabs
    isdir = os.path.isdir
    isfile = os.path.isfile
    islink = os.path.islink
    ismount = os.path.ismount

    if hasattr(os.path, 'samefile'):
        samefile = os.path.samefile

    getatime = os.path.getatime
    atime = property(
        getatime, None, None,
        """ Last access time of the file. """)

    getmtime = os.path.getmtime
    mtime = property(
        getmtime, None, None,
        """ Last-modified time of the file. """)

    if hasattr(os.path, 'getctime'):
        getctime = os.path.getctime
        ctime = property(
            getctime, None, None,
            """ Creation time of the file. """)

    getsize = os.path.getsize
    size = property(
        getsize, None, None,
        """ Size of the file, in bytes. """)

    if hasattr(os, 'access'):
        def access(self, mode):
            """ Return true if current user has access to this path.

            mode - One of the constants os.F_OK, os.R_OK, os.W_OK, os.X_OK
            """
            return os.access(self, mode)

    def stat(self):
        """ Perform a stat() system call on this path. """
        return os.stat(self)

    def lstat(self):
        """ Like path.stat(), but do not follow symbolic links. """
        return os.lstat(self)

    if hasattr(os, 'statvfs'):
        def statvfs(self):
            """ Perform a statvfs() system call on this path. """
            return os.statvfs(self)

    if hasattr(os, 'pathconf'):
        def pathconf(self, name):
            return os.pathconf(self, name)


    # --- Modifying operations on files and directories

    def utime(self, times):
        """ Set the access and modified times of this file. """
        os.utime(self, times)

    def chmod(self, mode):
        os.chmod(self, mode)

    if hasattr(os, 'chown'):
        def chown(self, uid, gid):
            os.chown(self, uid, gid)

    def rename(self, new):
        os.rename(self, new)

    def renames(self, new):
        os.renames(self, new)


    # --- Create/delete operations on directories

    def mkdir(self, mode=0777):
        os.mkdir(self, mode)

    def makedirs(self, mode=0777):
        os.makedirs(self, mode)

    def rmdir(self):
        os.rmdir(self)

    def removedirs(self):
        os.removedirs(self)


    # --- Modifying operations on files

    def touch(self):
        """ Set the access/modified times of this file to the current time.
        Create the file if it does not exist.
        """
        fd = os.open(self, os.O_WRONLY | os.O_CREAT, 0666)
        os.close(fd)
        os.utime(self, None)

    def remove(self):
        os.remove(self)

    def unlink(self):
        os.unlink(self)


    # --- Links

    if hasattr(os, 'link'):
        def link(self, newpath):
            """ Create a hard link at 'newpath', pointing to this file. """
            os.link(self, newpath)

    if hasattr(os, 'symlink'):
        def symlink(self, newlink):
            """ Create a symbolic link at 'newlink', pointing here. """
            os.symlink(self, newlink)

    if hasattr(os, 'readlink'):
        def readlink(self):
            """ Return the path to which this symbolic link points.

            The result may be an absolute or a relative path.
            """
            return path(os.readlink(self))

        def readlinkabs(self):
            """ Return the path to which this symbolic link points.

            The result is always an absolute path.
            """
            p = self.readlink()
            if p.isabs():
                return p
            else:
                return (self.parent / p).abspath()


    # --- High-level functions from shutil

    copyfile = shutil.copyfile
    copymode = shutil.copymode
    copystat = shutil.copystat
    copy = shutil.copy
    copy2 = shutil.copy2
    copytree = shutil.copytree
    if hasattr(shutil, 'move'):
        move = shutil.move
    rmtree = shutil.rmtree


    # --- Special stuff from os

    if hasattr(os, 'chroot'):
        def chroot(self):
            os.chroot(self)

    if hasattr(os, 'startfile'):
        def startfile(self):
            os.startfile(self)