<title>Python for .NET</title>

<style type="text/css"><!--

body {

font: 8pt/16pt georgia,verdana;

text-decoration: none;

color: #555753;

background: #ffffff;

margin: 0px;

}

p {

font: 8pt/16pt georgia;

}

h1 {

font: bold 14pt;

color: #000044;

}

h2 {

font: bold 14pt;

margin-bottom: 0px;

color: #000044;

}

h3 {

font: bold 12pt;

margin-bottom: 0px;

color: #000044;

}

a:link {

font-weight: bold;

text-decoration: none;

color: #000066;

}

a:visited {

font-weight: bold;

text-decoration: none;

color: #000066;

}

a:hover, a:active {

text-decoration: underline;

color: #000066;

}

pre {

font-family: monospace;

background-color: #EFEFFF;

color: #000044;

border-style: solid;

border-width: 1px;

border-color: #555753;

padding: 6px, 2px, 6px, 2px;

}

li {

font: 8pt/16pt georgia,verdana;

text-decoration: none;

color: #555753;

}

.spacer {

font: bold 14pt;

}

.menu {

background-color: #EFEFFF;

color: #000044;

border-style: solid;

border-width: 1px;

border-color: #555753;

padding: 6px, 2px, 6px, 2px;

font-size: x-small;

}

.menu ul {

margin: 0px;

padding: 0px;

}

//--></style>

<table width="98%" border="0" cellspacing="4">

<td align="left" valign="top" width="20%" class="menu">

<h1>Python for .NET</h1>

<ul>

<li><a href="#installation">Installation</a></li>

<li><a href="#getting_started">Getting Started</a></li>

<li><a href="#importing">Importing Modules</a></li>

<li><a href="#classes">Using Classes</a></li>

<li><a href="#fields">Fields and Properties</a></li>

<li><a href="#indexers">Using Indexers</a></li>

<li><a href="#methods">Using Methods</a></li>

<li><a href="#delegates">Delegates and Events</a></li>

<li><a href="#exceptions">Exception Handling</a></li>

<li><a href="#arrays">Using Arrays</a></li>

<li><a href="#collections">Using Collections</a></li>

<li><a href="#com">COM Components</a></li>

<li><a href="#types">Type Conversion</a></li>

<li><a href="#assemblies">Using Assemblies</a></li>

<li><a href="#embedding">Embedding Python</a></li>

<li><a href="#license">License</a></li>

</ul>

</td>

<td align="left" valign="top">

<p>

Python for .NET is a package that gives Python programmers nearly seamless

integration with the .NET Common Language Runtime (CLR) and provides a

powerful application scripting tool for .NET developers. Using this package

you can script .NET applications or build entire applications in Python,

using .NET services and components written in any language that targets the

CLR (Managed C++, C#, VB, JScript).

Note that this package does <em>not</em> implement Python as a first-class

CLR language - it does not produce managed code (IL) from Python code.

Rather, it is an integration of the C Python engine with the .NET runtime.

This approach allows you to use use CLR services and continue to use existing

Python code and C-based extensions while maintaining native execution speeds

for Python code. If you are interested in a pure managed-code implementation

of the Python language, you should check out the

<a href="http://www.ironpython.com">IronPython</a> project, which is in

active development.

Python for .NET is currently compatible with Python releases 2.3 and greater.

Current releases are available at the

<a href="http://pythonnet.sourceforge.net/">

Python for .NET website

</a>. To subscribe to the

<a href="http://mail.python.org/mailman/listinfo/pythondotnet">

Python for .NET mailing list

</a> or read the

<a href="http://mail.python.org/pipermail/pythondotnet/">

online archives

</a> of the list, see the

<a href="http://mail.python.org/mailman/listinfo/pythondotnet">

mailing list information

</a> page.

<a name="#installation"></a>

<h2>Installation</h2>

Python for .NET is available as a source release and as a Windows installer

for various versions of Python and the common language runtime from the

<a href="http://pythonnet.sourceforge.net/">

Python for .NET website

</a>. On Windows platforms, you can choose to install .NET-awareness into

an existing Python installation as well as install Python for .NET as a

standalone package.

<p>

The source release is a self-contained "private" assembly.

Just unzip the package wherever you want it, cd to that directory and run

python.exe to start using it. Note that the source release does not

include a copy of the CPython runtime, so you will need to have installed

Python on your machine before using the source release.

<strong>Running on Linux/Mono:</strong> preliminary testing shows that

PythonNet will run under <a href="http://www.go-mono.com">Mono</a>, though

the Mono runtime is not yet complete so there still may be problems. The

Python for .NET integration layer is 100% managed code, so there should be

no long-term issues under Mono - it should work better and better as the

Mono platform matures.

It is not currently possible to *build* PythonNet using only the Mono

tools, due to an issue involving the Mono assembler / disassembler. You

should, however, be able to try the pre-built assembly under Mono (or

compile it yourself with the MS tools and run it under Mono).

Note that if you are running under Mono on a *nix system, you will need

to have a compatible version of Python installed. You will also need

to create a symbolic link to the copy of libpython2.x.so (in your existing

Python installation) in the PythonNet directory. This is needed to ensure

that the mono interop dll loader will find it by name. For example:

ln -s /usr/lib/libpython2.4.so ./python24.so

<a name="getting_started"></a>

<h2>Getting Started</h2>

A key goal for this project has been that Python for .NET should &quot;work

just the way you'd expect in Python&quot;, except for cases that are .NET

specific (in which case the goal is to work &quot;just the way you'd expect

in C#&quot;).

If you already know Python, you can probably finish this readme and then

refer to .NET docs to figure out anything you need to do. Conversely if

you are familiar with C# or another .NET language, you probably just

need to pick up one of the many good Python books or read the Python

tutorial online to get started.

A good way to start is to run <strong>python.exe</strong> and follow along

with the examples in this document. If you get stuck, there are also a

number of demos and unit tests located in the source directory of the

distribution that can be helpful as examples.

<a name="importing"></a>

<h2>Importing Modules</h2>

<p>

Python for .NET allows CLR namespaces to be treated essentially as

Python packages. The top-level package is named <code>CLR</code>, and

acts as the root for accessing all CLR namespaces:

from CLR.System import String

import CLR.System as System

<p>

Types from any loaded assembly may be imported and used in this manner.

The import hook uses "implicit loading" to support automatic loading

of assemblies whose names correspond to an imported namespace:

# This will implicitly load the System.Windows.Forms assembly

from CLR.System.Windows.Forms import Form

Python for .NET uses the PYTHONPATH (sys.path) to look for assemblies

to load, in addition to the usual application base and the GAC. To

ensure that you can implicitly import an assembly, put the directory

containing the assembly in <code>sys.path</code>.

To load assemblies with names that do not correspond with a namespace,

you can use the standard mechanisms provided by the CLR:

from CLR.System.Reflection import Assembly

a = Assembly.LoadWithPartialName("SomeAssembly")

# now we can import namespaces defined in that assembly

from CLR.SomeNamespace import Something

Note that CLR modules are "lazy". Because a namespace can contain a

potentially very large number of classes, reflected CLR classes are

created on-demand when they are requested of a CLR module. This means

that using the Python <code>dir()</code> function in the interactive

interpreter will not necessarily show all of the classes available from

a given CLR module (it will only show any classes that have been referenced

to that point in time).

It also means that <code>from somemodule import *</code> will not work as

expected. It is possible that it could be made to work in the future,

but for now it would impose a big performance hit and a lot of

complexity to support something that is used relatively rarely and

often considered bad form in Python anyway ;)

<a name="classes"></a>

<h2>Using Classes</h2>

Python for .NET allows you to use any non-private classes, structs,

interfaces, enums or delegates from Python. To create an instance of

a managed class, you use the standard instantiation syntax, passing

a set of arguments that match one of its public constructors:

from CLR.System.Drawing import Point

p = Point(5, 5)

You can also subclass managed classes in Python. See the

<code>helloform.py</code> file in the <code>/demo</code> directory of the

distribution for a simple Windows Forms example that demonstrates

subclassing a managed class.

<a name="fields"></a>

<h2>Fields And Properties</h2>

You can get and set fields and properties of CLR objects just as if

they were regular attributes:

from CLR.System import Environment

name = Environment.MachineName

Environment.ExitCode = 1

<a name="indexers"></a>

<h2>Using Indexers</h2>

If a managed object implements one or more indexers, you can call

the indexer using standard Python indexing syntax:

from CLR.System.Collections import Hashtable

table = Hashtable()

table["key 1"] = "value 1"

Overloaded indexers are supported, using the same notation one

would use in C#:

items[0, 2]

items[0, 2, 3]

<a name="methods"></a>

<h2>Using Methods</h2>

Methods of CLR objects behave generally like normal Python methods.

Static methods may be called either through the class or through an

instance of the class. All public and protected methods of CLR objects

are accessible to Python:

from CLR.System import Environment

drives = Environment.GetLogicalDrives()

It is also possible to call managed methods <code>unbound</code> (passing the

instance as the first argument) just as with Python methods. This is

most often used to explicitly call methods of a base class.

<em>Note that there is one caveat related to calling unbound methods: it

is possible for a managed class to declare a static method and an

instance method with the same name. Since it is not possible for the

runtime to know the intent when such a method is called unbound, the

static method will always be called.</em>

The docstring of CLR a method (__doc__) can be used to view the

signature of the method, including overloads if the CLR method is

overloaded. You can also use the Python <code>help</code> method to inspect

a managed class:

from CLR.System import Environment

print Environment.GetFolderPath.__doc__

help(Environment)

<a name="delegates"></a>

<h2>Delegates And Events</h2>

Delegates defined in managed code can be implemented in Python. A

delegate type can be instantiated and passed a callable Python object

to get a delegate instance. The resulting delegate instance is a true

managed delegate that will invoke the given Python callable when it

is called:

def my_handler(source, args):

print 'my_handler called!'

# instantiate a delegate

d = AssemblyLoadEventHandler(my_handler)

# use it as an event handler

AppDomain.CurrentDomain.AssemblyLoad += d

Multicast delegates can be implemented by adding more callable objects

to a delegate instance:

d += self.method1

d += self.method2

d()

Events are treated as first-class objects in Python, and behave in

many ways like methods. Python callbacks can be registered with event

attributes, and an event can be called to fire the event.

Note that events support a convenience spelling similar to that used

in C#. You do not need to pass an explicitly instantiated delegate

instance to an event (though you can if you want). Events support the

<code>+=</code> and <code>-=</code> operators in a way very similar to

the C# idiom:

def handler(source, args):

print 'my_handler called!'

# register event handler

object.SomeEvent += handler

# unregister event handler

object.SomeEvent -= handler

# fire the event

result = object.SomeEvent(...)

<a name="exceptions"></a>

<h2>Exception Handling</h2>

You can raise and catch managed exceptions just the same as you would

pure-Python exceptions:

from CLR.System import NullReferenceException

try:

raise NullReferenceException("aiieee!")

except NullReferenceException, e:

print e.Message

print e.Source

<a name="arrays"></a>

<h2>Using Arrays</h2>

Managed arrays support the standard Python sequence protocols:

items = SomeObject.GetArray()

# Get first item

v = items[0]

items[0] = v

# Get last item

v = items[-1]

items[-1] = v

# Get length

l = len(items)

# Containment test

test = v in items

Multidimensional arrays support indexing using the same notation one

would use in C#:

items[0, 2]

items[0, 2, 3]

<a name="collections"></a>

<h2>Using Collections</h2>

Managed arrays and managed objects that implement the IEnumerable

interface can be iterated over using the standard iteration Python

idioms:

domain = System.AppDomain.CurrentDomain

for item in domain.GetAssemblies():

name = item.GetName()

<a name="com"></a>

<h2>Using COM Components</h2>

Using Microsoft-provided tools such as <strong>aximp.exe</strong> and

<strong>tlbimp.exe</strong>, it is possible to generate managed wrappers

for COM libraries. After generating such a wrapper, you can use the

libraries from Python just like any other managed code.

Note: currently you need to put the generated wrappers in the GAC,

in the PythonNet assembly directory or on the PYTHONPATH in order

to load them.

<a name="types"></a>

<h2>Type Conversion</h2>

Type conversion under Python for .NET is fairly straightforward - most

elemental Python types (string, int, long, etc.) convert automatically

to compatible managed equivalents (String, Int32, etc.) and vice-versa.

Note that all strings returned from the CLR are returned as unicode.

</p>

Types that do not have a logical equivalent in Python are exposed as

instances of managed classes or structs (System.Decimal is an example).

The .NET architecture makes a distinction between <code>value types</code>

and <code>reference types</code>. Reference types are allocated on the heap,

and value types are allocated either on the stack or in-line within an

object.

A process called <code>boxing</code> is used in .NET to allow code to treat

a value type as if it were a reference type. Boxing causes a separate

copy of the value type object to be created on the heap, which then

has reference type semantics.

Understanding boxing and the distinction between value types and

reference types can be important when using Python for .NET because

the Python language has no value type semantics or syntax - in

Python &quot;everything is a reference&quot;.

Here is a simple example that demonstrates an issue. If you are an

experienced C# programmer, you might write the following code:

items = CLR.System.Array.CreateInstance(Point, 3)

for i in range(3):

items[i] = Point(0, 0)

items[0].X = 1 # won't work!!

While the spelling of <code>items[0].X = 1</code> is the same in C# and

Python, there is an important and subtle semantic difference. In C# (and other

compiled-to-IL languages), the compiler knows that Point is a value

type and can do the Right Thing here, changing the value in place.

In Python however, "everything's a reference", and there is really no

spelling or semantic to allow it to do the right thing dynamically. The

specific reason that <code>items[0]</code> itself doesn't change is that

when you say <code>items[0]</code>, that getitem operation creates a Python

object that holds a reference to the object at <code>items[0]</code> via a

GCHandle. That causes a ValueType (like Point) to be boxed, so the following

setattr (<code>.X = 1</code>) <em>changes the state of the boxed value, not

the original unboxed value</em>.

The rule in Python is essentially: &quot;the result of any attribute or

item access is a boxed value&quot;, and that can be important in how you

approach your code.

Because there are no value type semantics or syntax in Python, you

may need to modify your approach. To revisit the previous example,

we can ensure that the changes we want to make to an array item

aren't "lost" by resetting an array member after making changes

to it:

items = CLR.System.Array.CreateInstance(Point, 3)

for i in range(3):

items[i] = Point(0, 0)

# This _will_ work. We get 'item' as a boxed copy of the Point

# object actually stored in the array. After making our changes

# we re-set the array item to update the bits in the array.

item = items[0]

item.X = 1

items[0] = item

This is not unlike some of the cases you can find in C# where you have

to know about boxing behavior to avoid similar kinds of <code>lost

update</code> problems (generally because an implicit boxing happened that

was not taken into account in the code).

This is the same thing, just the manifestation is a little different

in Python. See the .NET documentation for more details on boxing and

the differences between value types and reference types.

<a name="assemblies"></a>

<h2>Using Assemblies</h2>

The Python for .NET runtime uses Assembly.LoadWithPartialName to do

name-based imports, which will usually load the most recent version of

an assembly that it can find.

The CLR's ability to load different versions of assemblies side-by-side

is one of the (relatively few) places where the matching of meta-models

between Python and the CLR breaks down (though it is unclear how often

this happens in practice).

Because Python import is name-based and unaware of any concept of

versioned modules, the design goal has been for name-based implicit assembly

loading to do something consistent and reasonable (i.e. load most

recent available based on the name).

It <em>is</em> possible to load a specific version of an assembly if you

need to using code similar to the following:

from CLR.System.Reflection import Assembly, AssemblyName

name = AssemblyName(...) # set required version, etc.

assembly = Assembly.Load(name)

# now import namespaces from the loaded assembly

Things get a lot more complicated if you need to load more than one

version of a particular assembly (or more likely, you have a dependency

on some library the does so). In this case, the names you access via

the CLR modules will always come from the first version of the

assembly loaded (which will always win in the internals of the Python

for .NET runtime).

You <em>can</em> still use particular versions of objects in this case - you

just have to do more work to get the right versions of objects:

from CLR.System.Reflection import Assembly, AssemblyName

from System import Activator

name = AssemblyName(...) # get the right version

assembly = Assembly.Load(name)

type = assembly.GetType("QualifiedNameOf.TheTypeINeed")

obj = Activator.CreateInstance(type)

<a name="embedding"></a>

<h2>Embedding Python</h2>

<strong>Note:</strong> because Python code running under Python for .NET

is inherently unverifiable, it runs totally under the radar of the security

infrastructure of the CLR so you should restrict use of the Python assembly

to trusted code.

The Python runtime assembly defines a number of public classes that

provide a subset of the functionality provided by the Python C API.

These classes include PyObject, PyList, PyDict, etc. The source and

the unit tests are currently the only API documentation.. The rhythym

is very similar to using Python C++ wrapper solutions such as CXX.

At a very high level, to embed Python in your application you

will need to:

<li>Reference Python.Runtime.dll in your build environment</li>

<li>Call PythonEngine.Intialize() to initialize Python</li>

<li>Call PythonEngine.ImportModule(name) to import a module</li>

The module you import can either start working with your managed app

environment at the time its imported, or you can explicitly lookup and

call objects in a module you import.

For general-purpose information on embedding Python in applications, use

www.python.org or Google to find (C) examples. Because Python for .NET is

so closely integrated with the managed environment, you will generally be

better off importing a module and deferring to Python code as early as

possible rather than writing a lot of managed embedding code.

<strong>Important Note for embedders:</strong> Python is not free-threaded

and uses a global interpreter lock to allow multi-threaded applications to

interact safely with the Python interpreter. Much more information about

this is available in the Python C API documentation on the www.python.org

Website.

When embedding Python in a managed application, you have to manage the GIL

in just the same way you would when embedding Python in a C or C++

application.

Before interacting with any of the objects or APIs provided by the

Python.Runtime namespace, calling code must have acquired the Python

global interpreter lock by calling the <code>PythonEngine.AcquireLock</code>

method. The only exception to this rule is the

<code>PythonEngine.Initialize</code> method, which may be called at startup

without having acquired the GIL.

When finished using Python APIs, managed code must call a corresponding

<code>PythonEngine.ReleaseLock</code> to release the GIL and allow other

threads to use Python.

The AcquireLock and ReleaseLock methods are thin wrappers over the

unmanaged <code>PyGILState_Ensure</code> and <code>PyGILState_Release</code>

functions from the Python API, and the documentation for those APIs applies

to the managed versions.

<a name="license" />

<h2>License</h2>

<p>

Python for .NET is released under the open source Zope Public License (ZPL).

A copy of the ZPL is included in the distribution, or you can find a copy

of the

<a href="http://pythonnet.sourceforge.net/license.txt">

ZPL online

</a>. Some distributions of this package include a copy of the C Python

dlls and standard library, which are covered by the

<a href="http://www.python.org/license.html">

Python license

</a>.

</td>