The ‘py’ lib provides a uniform high-level api to deal with filesystems and filesystem-like interfaces: py.path. It aims to offer a central object to fs-like object trees (reading from and writing to files, adding files/directories, examining the types and structure, etc.), and out-of-the-box provides a number of implementations of this API.

py.test.local - local file system path

basic interactive example

The first and most obvious of the implementations is a wrapper around a local filesystem. It’s just a bit nicer in usage than the regular Python APIs, and of course all the functionality is bundled together rather than spread over a number of modules.

Example usage, here we use the py.test.ensuretemp() function to create a py.path.local object for us (which wraps a directory):

>>> import py
>>> temppath = py.test.ensuretemp('py.path_documentation')
>>> foopath = temppath.join('foo') # get child 'foo' (lazily)
>>> foopath.check() # check if child 'foo' exists
>>> foopath.write('bar') # write some data to it
>>> foopath.check()
>>> foofile = # return a 'real' file object

reference documentation

py.path.svnurl and py.path.svnwc

Two other py.path implementations that the py lib provides wrap the popular Subversion revision control system: the first (called ‘svnurl’) by interfacing with a remote server, the second by wrapping a local checkout. Both allow you to access relatively advanced features such as metadata and versioning, and both in a way more user-friendly manner than existing other solutions.

Some example usage of py.path.svnurl:

.. >>> import py
.. >>> if not py.test.config.option.urlcheck: raise ValueError('skipchunk')
>>> url = py.path.svnurl('')
>>> info =
>>> info.kind
>>> firstentry = url.log()[-1]
>>> import time
>>> time.strftime('%Y-%m-%d', time.gmtime(

Example usage of py.path.svnwc:

.. >>> if not py.test.config.option.urlcheck: raise ValueError('skipchunk')
>>> temp = py.test.ensuretemp('py.path_documentation')
>>> wc = py.path.svnwc(temp.join('svnwc'))
>>> wc.checkout('')
>>> wc.join('').check()

Common vs. specific API, Examples

All Path objects support a common set of operations, suitable for many use cases and allowing to transparently switch the path object within an application (e.g. from “local” to “svnwc”). The common set includes functions such as to read all data from a file, path.write() to write data, path.listdir() to get a list of directory entries, path.check() to check if a node exists and is of a particular type, path.join() to get to a (grand)child, path.visit() to recursively walk through a node’s children, etc. Only things that are not common on ‘normal’ filesystems (yet), such as handling metadata (e.g. the Subversion “properties”) require using specific APIs.

A quick ‘cookbook’ of small examples that will be useful ‘in real life’, which also presents parts of the ‘common’ API, and shows some non-common methods:

Searching .txt files

Search for a particular string inside all files with a .txt extension in a specific directory.

>>> dirpath = temppath.ensure('testdir', dir=True)
>>> dirpath.join('textfile1.txt').write('foo bar baz')
>>> dirpath.join('textfile2.txt').write('frob bar spam eggs')
>>> subdir = dirpath.ensure('subdir', dir=True)
>>> subdir.join('textfile1.txt').write('foo baz')
>>> subdir.join('textfile2.txt').write('spam eggs spam foo bar spam')
>>> results = []
>>> for fpath in dirpath.visit('*.txt'):
...     if 'bar' in
...         results.append(fpath.basename)
>>> results.sort()
>>> results
['textfile1.txt', 'textfile2.txt', 'textfile2.txt']

Working with Paths

This example shows the py.path features to deal with filesystem paths Note that the filesystem is never touched, all operations are performed on a string level (so the paths don’t have to exist, either):

>>> p1 = py.path.local('/foo/bar')
>>> p2 = p1.join('baz/qux')
>>> p2 == py.path.local('/foo/bar/baz/qux')
>>> sep = py.path.local.sep
>>> p2.relto(p1).replace(sep, '/') # os-specific path sep in the string
>>> p2.bestrelpath(p1).replace(sep, '/')
>>> p2.join(p2.bestrelpath(p1)) == p1
>>> p3 = p1 / 'baz/qux' # the / operator allows joining, too
>>> p2 == p3
>>> p4 = p1 + ".py"
>>> p4.basename == ""
>>> p4.ext == ".py"
>>> p4.purebasename == "bar"

This should be possible on every implementation of py.path, so regardless of whether the implementation wraps a UNIX filesystem, a Windows one, or a database or object tree, these functions should be available (each with their own notion of path seperators and dealing with conversions, etc.).

Checking path types

Now we will show a bit about the powerful ‘check()’ method on paths, which allows you to check whether a file exists, what type it is, etc.:

>>> file1 = temppath.join('file1')
>>> file1.check() # does it exist?
>>> file1 = file1.ensure(file=True) # 'touch' the file
>>> file1.check()
>>> file1.check(dir=True) # is it a dir?
>>> file1.check(file=True) # or a file?
>>> file1.check(ext='.txt') # check the extension
>>> textfile = temppath.ensure('text.txt', file=True)
>>> textfile.check(ext='.txt')
>>> file1.check(basename='file1') # we can use all the path's properties here

Setting svn-properties

As an example of ‘uncommon’ methods, we’ll show how to read and write properties in an py.path.svnwc instance:

.. >>> if not py.test.config.option.urlcheck: raise ValueError('skipchunk')
>>> wc.propget('foo')
>>> wc.propset('foo', 'bar')
>>> wc.propget('foo')
>>> len(wc.status().prop_modified) # our own props
>>> msg = wc.revert() # roll back our changes
>>> len(wc.status().prop_modified)

SVN authentication

Some uncommon functionality can also be provided as extensions, such as SVN authentication:

.. >>> if not py.test.config.option.urlcheck: raise ValueError('skipchunk')
>>> auth = py.path.SvnAuth('anonymous', 'user', cache_auth=False,
...             interactive=False)
>>> wc.auth = auth
>>> wc.update() # this should work
>>> path = wc.ensure('thisshouldnotexist.txt')
>>> try:
...     path.commit('testing')
... except py.process.cmdexec.Error, e:
...     pass
>>> 'authorization failed' in str(e)

Known problems / limitations

  • The SVN path objects require the “svn” command line, there is currently no support for python bindings. Parsing the svn output can lead to problems, particularly regarding if you have a non-english “locales” setting.
  • While the path objects basically work on windows, there is no attention yet on making unicode paths work or deal with the famous “8.3” filename issues.