Just hailed like hell here, thought a window was going to break ;o)

Well, I've started working with the latest drop of Base4.net (version 2.1) - in fact I've decided to take the plunge and actually start using it for the project I'm currently working on (a work project, rather then a home project) so I should be giving it a pretty good thrashing over the next 3-6 months.  I already had a rough schema in place, so I generated types from it using the bundled web admin... It seems pretty cool... though I've struck a few things:

• nvarchar(256) fields seems to show up as being 512 characters in length within base4.net...  I'm not sure if this is a bug with the schema interrogation, guess we'll find out as a I dig further in ;o) - though it's only skin deep of course, as you can change them to 256 with a quick edit.
  
• I managed to create an assembly that was invalid... not sure how ;o) I could look at in reflector .Net, but base4 wasn't having a bar of it (even though it had generated it) - I must have screwed up a relationship somewhere, but I wasn't really sure... ended up just dumping it and starting from scratch again.
  
• It doesn't seem that easy to refactor existing schema's using the web designer... while I was learning it would have been good to generate a schema, play with the assembly, then start editing and generate again... instead I ended up having to save the schema to disk, delete, edit, and recreate.  Still, it's pretty slick :)
  

I think though it probably all comes down to:

• My urge to make things happen without understanding the in's and outs of base4.net
• Horses for courses, the web admin is great for getting off to ground with base4.net, but I think I'll have to get down and dirty with the xml pretty soon...
  

Though I can't discuss the project, I can discuss what I'll be building it in... so we have:

• Base4.Net
• SQL 2005
• Castle (Monorail & IoC)
• NLog (or log4net... though I'm loosing faith in the latter)
• NUnit
• RhinoMocks
• Splicer.Net
• Some COM & Unmanaged code for audio processing

This is also the first time I've used monorail for a commercial project (we already use a lot of the facilities and the windsor container in the Syzmk RMP product) so it's certainly some interesting times... as we progress I'll give some feedback on both the Monorail and Base4.Net "user experience" :)

Back again!

Here is the source for the last post: ironpython-and-nunit.zip (23.01 KB) - it's pretty rough around the edges (and in the middle :P) because I was just trying to see if it was possible, rather then to produce some production-level toolkit for integrating IronPython and NUnit.

Edit 2008-03-14: An updated version of the code target NUnit 2.4 and VS2008 can be found here.

Usage

First off... usage is pretty straightforward.

• Test fixtures must have NUnitFixture as one of their base classes.
• Test methods must be prefixed with "test".
• The setup method, if you need one, must be called "setUp"
• The teardown method, if you need one, must be called "tearDown"
• Fixture setup and teardown methods are also supported, they are called "setUpFixture" and "tearDownFixture".
• Assert.XXX methods are the same as NUnit, unless you want to assert an exception is thrown.
• For asserting an exception is thrown, make a call to "self.failUnlessRaises(...)" - it has the same syntax as the PyUnit equivalent, however it's tweaked slightly so it will work with Python exceptions, or CLR exceptions.
• Do not perform an "Import * from System" in your test fixtures - it will clober the native Python Exception type, and may give you odd behaviour.
  

Implementation

There are 8 classes and one python script which make up the implementation, the 8 classes are split between the "model" (basically the python-side of the equation and the "NUnit extension" which are the necessary extensions to expose the model to NUnit...) - let's run through it quickly:

First we have the AbstractPythonSuite which hosts the python engine, and handles configuring and running the scripts containing the test fixtures.  Then we have a couple of classes which form the "model" consisting of a PythonFixture, representing a fixture, and PythonTestCase, representing a single test method in the fixture.

To build the model we have a PythonFixtureBuilder which is passed the python engine (from within the AbstractPythonSuite) and spits out all the fixture models, with the test cases assigned... this also assigns the various delegates for the setup/teardown/setupFixture/teardownFixture methods.

At this point the model is complete, it's just a matter of integrating it with NUnit - for this we have some more classes.

First of we need the PythonSuiteExtension, which represents the overall test suite, this contains PythonFixtureExtension's, for the fixtures, and finally PythonTestMethod for the individual test methods - the python extensions take care of running the tests in the model and calling the setup and tear down methods at the right times, and recording the results of course.

Last of all we have the PythonSuiteExtensionBuilder as mentioned in the last post, which takes care of creating the PythonSuiteExtension's for each class derived from AbstractPythonSuite it discovers in the target assembly.

Last of all we have a small python script which provides the necessary implementation for the base NUnitFixture class that all test fixtures are derived from... and a helper method we use in the PythonFixtureBuilder to identify the test methods in a class.


And that's all there is to it...  If anyone actually finds this useful and wants to build on from it, please feel free, and if at all possible keep it open source (though you don't have to...)

Something to keep in mind is that at the moment if your test fixture can't actually be parsed and executed by the python engine, that exception will bubble up, eventually stopping the test assembly from being loaded at all... I think the nicer solution to this would be to add a custom NUnit test case for each script that failed to load, which in turn fails when executed as part of the assembly - dumping out the exception raised during construction.

At least then a tool like TestDriven.Net would quickly notify the developer of their mistake, instead of just falling over.

And very last of all, though I've been manually specifying the names of the script files.. using a class declaration like this:


You could make it a little more dynamic and do something like this:


Which will load all scripts with "#test" in the first line, which seem'is a little easier to me (not having to spell out the path to the script, and only having to create the test file to make things happen) - YMMV of course.

Breaking scripts

I’ve been thinking about the issues involved with using IronPython in a product as a scripting language, internally we’ve already struggled at Syzmk with my love of refactoring breaking scripts in our deployed products left right and centre.

I think the problems I need to solve first are:

• Ensuring an understanding of the user experience when scripting with the application (read: dogfooding the scripting experience)
  
• Making sure I identify where helper classes / wrappers etc. may be necessary to make the scripting experience more palatable to end users.
• Ensuring any changes don’t clobber the contracts between the scripting language and the client when refactoring the underlying .Net classes they rely on.

Regression tests for expected script usage is what it comes down… but if you sit down and start trying to write these tests in NUnit it all gets a bit unwieldy – there’s a lot of time wasted making sure your tabs are right, escaping characters correctly in strings, and running scripts, evaluating the results, and then asserting against them – it's frustrating, and fails to emulate the experience of your users who are writing scripts purely in Python.

PyUnit?

So my next port of call was to try the unit testing support that comes with python (PyUnit) – it would definitely do the job – but then I need to do a couple of things… first I need to start bundling quite a few standard libraries with the scripting engine in our product, and then I have to write additional functionality to support capturing the results of running the PyUnit tests and presenting them to the automated build process… It's all sounding like a pain in the ass, and  the automation build works nicely as it is, I’d rather not have to mess with it...

So I decided to do a little integration work, to see if I could run a test suite written in python as part of a C# assembly.

NUnit meets (Iron)Python

First off, my solutions user experience is ok – not perfect – but then I just wanted to make sure I could get it working, before I put any effort into cleaning it up, and not being an expert in extending NUnit, or coding python, it was probably doomed from the start to be a mediocre solution ;o)

The steps to get it working in a test assembly are:

1. Writing a fixture in python, and including it in the assembly as an embedded resource.
2. Creating a suite which references the embedded python files - this is a class derived from the "AbstractPythonSuite".
   
3. Creating a Suite builder (I had to do this in the test assembly itself, because it wasn’t working when I bundled this in a support library).
4. Run the suite ;o)

First off, we write the fixture, the only requirements are:

• The fixture class is derived from “NUnitFixture”.
• The test methods are prefixed with “test” – as per PyUnit. And take only a “self” argument.

Here’s an example fixture:

As you can see we are just using the existing NUnit Assert methods… There is one caveat to this that I will cover in another post (that of testing for expected exceptions).

Next we create a suite – which can wrap one or more python script files (included as assembly resources) – this is all we need to bridge the gap.

And last of all, we need to drop in a suite builder – in theory I should be able to include this in a support assembly – but it doesn’t seem to work, here’s the code:

At this point when can test the assembly with NUnitGui, and see our python fixtures appear under the “MyPythonSuite”… pretty slick huh?

The big downsides of this approach are that you can’t run the tests individually using something like TestDriven.Net, and Resharper’s unit test window doesn’t pick them up… However NUnitGui and the corresponding NUnitConsole have no problems with them, and running all the tests in your assembly with TestDriven.Net does pick them up thankfully – so as long you don’t have many slow tests in your assembly it should still be tolerable (you can run them while still coding...).

And of course, they will be picked up in your automated builds – which is great news for me... I hate messing around with CruiseControl.Net's configuration.

I’ll tidy up the implementation a little and post the code soon.

Well, Seo Sanghyeon asked me a pretty reasonable question of the last entry

"Why don't you use help(), instead of printing Overloads?"

Good question, I didn't realise that actually worked in IronPython for .Net types, I was very pleasantly surprised, I remember in the early betas i tried using help() on various managed types and it would just raise a NotImplementedException exception (or something like that, I forget now) - so what does that mean for my examples... well... let's take a system type like Guid and use help() on it.

IronPython 1.0.60816 on .NET 2.0.50727.42
Copyright (c) Microsoft Corporation. All rights reserved.
>>> import System
>>> from System import *
>>> help(Guid)
Help on Guid in module System in mscorlib, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089

 |      Guid(Array[Byte] b)
 |      Guid(UInt32 a, UInt16 b, UInt16 c, Byte d, Byte e, Byte f, Byte g, Byteh, Byte i, Byte j, Byte k)
 |      Guid(str g)
 |      Guid(int a, Int16 b, Int16 c, Array[Byte] d)
 |      Guid(int a, Int16 b, Int16 c, Byte d, Byte e, Byte f, Byte g, Byte h, Byte i, Byte j, Byte k)
 |      Guid CreateInstance[Guid]()
 |
 |  Data and other attributes defined here:
 |
 |      CompareTo(...)
 |              int CompareTo(self, object value)
 |              int CompareTo(self, Guid value)
 |      Equals(...)
 |              bool Equals(self, object o)
 |              bool Equals(self, Guid g)
 |      Finalize(...)
 |              Finalize(self)
 |      GetHashCode(...)
 |              int GetHashCode(self)
 |      GetType(...)
 |              Type GetType(self)
 |      MemberwiseClone(...)
 |              object MemberwiseClone(self)
 |      NewGuid(...)
 |              Guid NewGuid()
 |      ToByteArray(...)
 |              Array[Byte] ToByteArray(self)
 |      ToString(...)
 |              str ToString(self)
 |              str ToString(self, str format)
 |              str ToString(self, str format, IFormatProvider provider)
 |      __eq__(...)
 |              bool op_Equality(Guid a, Guid b)
 |      __init__(...)
 |              x.__init__(...) initializes x; see x.__class__.__doc__ for signature
 |              x.__init__(...) initializes x; see x.__class__.__doc__ for signature
 |      __ne__(...)
 |              bool op_Inequality(Guid a, Guid b)
 |      __new__(...)
 |              __new__(cls, Array[Byte] b)
 |              __new__(cls, UInt32 a, UInt16 b, UInt16 c, Byte d, Byte e, Bytef, Byte g, Byte h, Byte i, Byte j, Byte k)
 |              __new__(cls, str g)
 |              __new__(cls, int a, Int16 b, Int16 c, Array[Byte] d)
 |              __new__(cls, int a, Int16 b, Int16 c, Byte d, Byte e, Byte f, Byte g, Byte h, Byte i, Byte j, Byte k)
 |              Guid CreateInstance[Guid]()
 |

>>>

Sweet, it works a treat - and will save a lot of digging time - though you might want to capture the results for big classes ;o)

Thanks Seo!

(btw, Seo is the man behind the Iron Python Community Edition & FePy projects - which can be found on sourceforge) and I would have to agree with this article that he does seem almost OmniPresent in the IronPython community...

This entry was going to cover thinking about your .Net classes from the perspective of Python consumption… but actually it's not, we'll cover that another time.

What this post is really going to be about is how to spelunk a little into our managed classes via the IronPython interactive console - and to see how some common practices in .Net have interesting results in python.

This entry is going to be a little slow paced, I’m making the assumption you’re a .Net developer, and haven’t cut Python code before, if you know your way around Python probably best to pass this one over.

Also worth noting at this point that I'm really just a Python hacker, and certainly no guru...

So first off, we have a class with a bunch of overloads for various methods, here is the interface… I’ve left the code for the class itself out, but the class is called MyFileStore.

Pretty simple, now lets look at using this class with Iron python, I’ll be running with the ipy (the interactive interpreter) so we can investigate our types as we go along (this is an executable which comes with the Iron Python binaries). The ipy console has a prompt consisting of three greater then signs “>>>” so you’ll know what I’m typing into the console by looking for those.

First off, we need to load our type… this is pretty trivial…

>>> import clr
>>> clr.AddReferenceToFileAndPath("D:\\Demo\\SampleIronPythonLib.dll")
>>> from IronPythonLib import *

All done, lets make sure our class is in there, using the inbuilt function dir() – a list of built-in functions can be found here: http://www.python.net/quick-ref1_52.html#BuiltIn

>>> dir()
['IMyFileStore', 'MyFileStore', '_', '__builtins__', '__doc__', '__name__', 'clr', 'results', 'site', 'store', 'sys']

MyFileStore, sweet.

First off, lets create an instance of my type (not yours ;o)

>>> store = new MyFileStore()
Traceback (most recent call last):
SyntaxError: unexpected token MyFileStore (<stdin>, line 1)
>>>

Arse, python doesn’t have a new operator… forgot that, lets try it again.

>>> store = MyFileStore

Right, looks good, interpreter didn’t return any errors… but wait, lets make sure we got what we want, which is an instance of MyFileStore….

>>> store
<type ‘MyFileStore’>
>>>

Woops, that’s not what we wanted, we want an instance, not the type… lets try again:

>>> store = MyFileStore()
>>> store
<MyFileStore object at 0x0000000000000035>
>>>

That’s better, lets find out what our store is capable off, using the dir command.

>>> dir(store)
['AddFile', 'AddFiles', 'Equals', 'ExtractFiles', 'Finalize', 'GetHashCode', 'GetType', 'MakeDynamicType', 'MemberwiseClone', 'ReadFile', 'Reduce', 'ReferenceEquals', 'ToString', '__class__', '__doc__', '__init__', '__module__', '__new__','__reduce__', '__reduce_ex__', '__repr__']
>>>

We see our AddFile, AddFiles,ReadFile & ExtractFiles methods – everything looks to be in order… but what are the parameters for those methods?

>>> dir(store.AddFile)
['Call', 'CallInstance', 'Equals', 'Finalize', 'GetHashCode', 'GetTargetType', '
GetType', 'Make', 'MakeDelegate', 'MakeDynamicType', 'MemberwiseClone', 'Overloads', 'PrependInstance', 'Reduce', 'ReferenceEquals', 'ToString', '__call__', '__class__', '__doc__', '__eq__', '__getitem__', '__hash__', '__init__', '__module__', '__name__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__self__', '__str__']
>>>

Hmm… what’s the “Overloads” thing, lets find out…

>>> dir(store.AddFile.Overloads)
['Equals', 'Finalize', 'Function', 'GetHashCode', 'GetOverload', 'GetTargetFunction', 'GetType', 'MakeDynamicType', 'MemberwiseClone', 'Reduce', 'ReferenceEquals', 'Targets', 'ToString', '__class__', '__doc__', '__getitem__', '__init__', '__module__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__str__']
>>>

That doesn’t seem very useful, though what if we check it’s string representation?

>>> store.AddFile.Overloads
<IronPython.Runtime.Calls.BuiltinFunctionOverloadMapper object at 0x0000000000000036 [{'Guid AddFile(self, str fileName)': <built-in function AddFile>, 'Guid AddFile(self, str fileName, str mimeType)': <built-in function AddFile>, 'Guid AddFile(self, Array[Byte] contents, str mimeType)': <built-in function AddFile>, 'Guid AddFile(self, Array[Byte] contents)': <built-in function AddFile>}]>
>>>

Look at that, there are our 4 overloads – but it’s a bit jumbled, what am I looking at it….

Well at first glance it looks a bit like a python dictionary… notice the colon separating the signature for the method and <built-in function AddFile> … what if we try to grab one for use later, maybe for some functional programming, why don’t we get the one which takes the fileName and mimeType parameters eh?

>>> myFunc = store.AddFile.Overloads[str,str]
Traceback (most recent call last):
File , line 0, in <stdin>##100
TypeError: __getitem__() takes exactly 1 argument (1 given)
>>>

Bugger, that didn’t work… still, not a bad guess - one parameter eh? Lets have another stab, maybe it just wants us to group the types together somehow?

>>> myFunc = store.AddFile.Overloads[(str,str)]
>>> myFunc
<built-in method AddFile of MyFileStore object at 0x0000000000000035>
>>>

Looks better, of course if we wanted to we could use our function right now… let’s give it a whirl:

>>> myFunc("d:\\testinput.3gp", "video/3gpp")
<System.Guid object at 0x0000000000000037 [7b6d6dc5-bbe6-4831-bff7-21984031684f]>
>>>

Looks to have worked, the method has successfully returned us a Guid identifying the new file… shame we didn’t save it into a variable for future reference, cest la vie.

Luckily I can just copy and paste it – lets see if I can get the file back using this overload…

Simple as.. lets try…

>>> id = Guid("7b6d6dc5-bbe6-4831-bff7-21984031684f")
Traceback (most recent call last):
File , line 0, in <stdin>##116
NameError: name 'Guid' not defined
>>>

Woops, we need to import the System namespace…

>>> import System

Now lets try again…

>>> id = Guid("7b6d6dc5-bbe6-4831-bff7-21984031684f")
Traceback (most recent call last):
File , line 0, in <stdin>##116
NameError: name 'Guid' not defined
>>>

Huh… oh wait, let’s check our local symbol table:

>>> dir()
['MyFileStore', 'System', '_', '__builtins__', '__doc__', '__name__', 'clr', 'myArray', 'myFunc', 'results', 'site', 'store', 'sys']

System eh? Lets dig into it…

>>> dir(System)
['AccessViolationException', 'Action', 'ActivationContext', 'Activator', 'AppDom
ain', 'AppDomainInitializer', 'AppDomainManager', 'AppDomainManagerInitializatio
nOptions', 'AppDomainSetup', 'AppDomainUnloadedException', 'ApplicationException… and all the rest.
>>

Well, what can we do about that… we could import them all using “from System import *” – but we only need Guid, so why don’t we just grab that eh?

>>> from System import Guid
>>> dir()
['Guid', 'MyFileStore', 'System', '_', '__builtins__', '__doc__', '__name__', 'clr', 'myArray', 'myFunc', 'results', 'site', 'store', 'sys']
>>>

Sweet, now lets try it again… finally!

>>> id = Guid("7b6d6dc5-bbe6-4831-bff7-21984031684f")
>>> mimeType = ""
>>> contents = store.ReadFile(id, mimeType)
Traceback (most recent call last):
File , line 0, in <stdin>##136
File , line 0, in ReadFile##66
File D:\dev\Projects\IronPythonDemo\IronPythonLib\MyClass.cs, line 70, in Read
File
File mscorlib, line unknown, in WriteAllBytes
File mscorlib, line unknown, in .ctor
File mscorlib, line unknown, in Init
ValueError: Empty path name is not legal.
>>>

Oh no! what went wrong??… what’s this about an empty path name, I was expecting to get some bytes back…. Hmmm, don’t panic, lets just do some digging…

>>> store.ReadFile.Overloads
<IronPython.Runtime.Calls.BuiltinFunctionOverloadMapper object at 0x0000000000000039 [{'(Array[Byte], str) ReadFile(self, Guid id)': <built-in function ReadFile>, 'ReadFile(self, Guid id, str fileName)': <built-in function ReadFile>}]>
>>>

Well there’s the problem, I’ve called the wrong version, but… that’s odd, there’s a ReadFile there I haven’t defined in my class… and one that’s missing… hmmm, time to file a bug report.

Or not, this is just where the Python and .Net world don’t see eye to eye – Python doesn’t know about out parameters as such, so it simulates the effect by altering the method’s signature – lets have a go at playing by IronPython's rules:

>>> contentsAndMimeType = store.ReadFile(id)
>>> len(contentsAndMimeType)
2
>>> type(contentsAndMimeType[0])
<type 'Array[Byte]'>
>>> type(contentsAndMimeType[1])
<type 'str'>
>>> contentsAndMimeType[1]
'video/3gpp'
>>>

So our “out string mimeType” parameter was returned by the method instead, as the second item in an array… hmm… interesting.

Well if python does this to out parameters, what does it do to ref parameters – lets try the extract files method to see what happens… this method is implemented to create a new list if we don’t supply a valid instance, but lets pass in a valid one first – I wont be caught out this time, so I’ll import the types I need first (List<T>):

>>> from System.Collections.Generic import *
>>> bytes = List<byte[]>()
Traceback (most recent call last):
SyntaxError: unexpected token ] (<stdin>, line 1)
>>>

Oh dear, looks like Python doesn’t use the <T> syntax… lets start digging again…

>>> dir(List)
['Add', 'AddRange', 'AsReadOnly', 'BinarySearch', 'Capacity', 'Clear', 'Contains
', 'ConvertAll', 'CopyTo', 'Count', 'Enumerator', 'Equals', 'Exists', 'Finalize'
, 'Find', 'FindAll', 'FindIndex', 'FindLast', 'FindLastIndex', 'ForEach', 'GetEn
umerator', 'GetHashCode', 'GetRange', 'GetType', 'IndexOf', 'Insert', 'InsertRange', 'LastIndexOf', 'MakeDynamicType', 'MemberwiseClone', 'Reduce', 'ReferenceEquals', 'Remove', 'RemoveAll', 'RemoveAt', 'RemoveRange', 'Reverse', 'Sort', 'ToArray', 'ToString', 'TrimExcess', 'TrueForAll', '__class__', '__doc__', '__getitem__', '__init__', '__module__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setitem__']
>>> List
<type 'List[T]'>
>>>

Hmm… nothing obvious there, but we could hazard a guess…

>>> List[str]
<type 'List[str]'>
>>>

Hmmm… so lets give our List of byte array another go…

>>> bytes = List[byte[]]()
Traceback (most recent call last):
SyntaxError: unexpected token ] (<stdin>, line 1)

Hmm… it doesn’t like byte, lets have a look…

>>> byte
Traceback (most recent call last):
File , line 0, in <stdin>##56
NameError: name 'byte' not defined
>>> Byte
<type 'Byte'>
>>>

Ahhh… I forgot that “byte” (lowercase) isn’t actually what the Clr calls it… hmm… we could try it again, but it wont work – simply because Byte[] doesn’t make sense (if we’re using square brackets to index into a generic type… ) so having another dig in the IronPython docs we see that we need to index to the type “Array” with the type of our array, given this knowledge we can get what we want…

>>> bytes = List[Array[Byte]]()
>>> bytes
<List[Array[Byte]] object at 0x000000000000002C>
>>>

Sweet, now to call that ExtractFiles method…

>>> store.ExtractFiles(bytes)
<List[Array[Byte]] object at 0x000000000000002C>
>>> bytes
<List[Array[Byte]] object at 0x000000000000002C>
>>> len(bytes)
8

Well it returned a list… but it did accept our reference parameter, and it has updated the collection we passed to it… but what if we had passed it a null instead?

>>> bytes = None
>>> store.ExtractFiles(bytes)
<List[Array[Byte]] object at 0x0000000000000030>
>>> len(bytes)
Traceback (most recent call last):
File , line 0, in <stdin>##83
File , line 0, in Length##75
TypeError: len() of unsized object of type <type 'NoneType'>
>>>

Hmmm… it returned a list, but it didn’t set the value in our variable "bytes" – this isn’t really ref like behavior… yet, if you think about it we can get almost the same thing by doing this:

>>> bytes = store.ExtractFiles(bytes)
>>> len(bytes)
8
>>>

Right, our journey has almost come to an end… looking at the interface we have a method declared like so:

params is handy in .Net code – anything to save tedious typing of array declarations, lets see if we can do the same thing in python?

>>> store.AddFiles("d:\\testinput1.3gp", "d:\\testinput2.3gp")
System.Guid[](<System.Guid object at 0x0000000000000031 [d3e8f099-3005-461c-a63a-fdc69b2091ee]>, <System.Guid object at 0x0000000000000032 [d9d1fb25-fc7a-47a2-89ac-c11264bfa47f]>)
>>>

Sweet, that’s a pleasant surprise after that whole out and ref debacle, I was beginning to loosing hope!

But wait, why not revive our faith in Python a little more by looking at a few tricks…

>>> paramsDict = { "fileName" : "d:\\testinput.3gp", "mimeType" : "video/3gpp"}
>>> store.AddFile(**paramsDict)
<System.Guid object at 0x0000000000000033 [e07680d9-f544-4847-9a1e-d04a0ef137f7]>
>>>

What did we just do? Well… given a dictionary, where the keys are the parameter names, we used them as the parameters for one of our .Net methods using the double asterisk syntax – think of what you’d have to do to code this in .Net… reflection city ;o)

It doesn’t just have to be dictionaries… we can use arrays too…

>>> paramsArray = [ "d:\\testinput.3gp", "video/3gpp" ]
>>> store.AddFile(*paramsArray)
<System.Guid object at 0x0000000000000034 [85006c43-e6f5-4e90-b837-0868788cf453]>
>>>

Neat, of course, being a dynamic language and all, the selection of the appropriate overload is based on the types in the array or dictionary… it would be a nightmare to do both of these in .Net with reflection yourself.

Faith still not restored eh, what about creating a python class that wraps an existing (possibly sealed) .Net class, that forwards calls on unless we want to override the behavior…

>>> class MyFileStoreWrapper:
... def __init__(self, realStore):
... self.realStore = realStore
... def ExtractFiles(self, myList):
... raise Exception, "this method isn't allowed"
... def __getattr__(self, name):
... if name == "ExtractFiles": return ExtractFiles
... return eval("self.realStore."+name)
...
>>> store = MyFileStoreWrapper(MyFileStore())
>>> store.AddFile("d:\\testinput.3gp")
<System.Guid object at 0x000000000000002C [a2bf8747-cc71-4b79-8c02-0b61a87ff67f]
>
>>> store.ExtractFiles(None)
Traceback (most recent call last):
File , line 0, in <stdin>##52
File , line 0, in ExtractFiles
Exception: this method isn't allowed
>>>

ExtractFiles can’t be invoked on our wrapper, however our wrapper automagically responds for other methods like AddFile – sadly I’ve used an eval (which, rhymes with Evil - coincidence?) here because MyFileStore doesn’t expose a __getattr__ method (because it's a .Net classes, not a native Python one).

The nice thing now, is that given a native python class, we can do some things we aren’t allowed to do to the underlying .Net class – like adding new methods at run time.

>>> def ExtractMoreFiles(self, myList):
... raise Exception, 'not implemented yet'
...
>>> MyFileStoreWrapper.ExtractMoreFiles = ExtractMoreFiles
>>> store.ExtractMoreFiles(None)
Traceback (most recent call last):
File , line 0, in <stdin>##60
File , line 0, in ExtractMoreFiles
Exception: not implemented yet
>>>

I think it’s pretty cool, and though python's metaprogramming model isn’t quite up to ruby standards, it’s still pretty easy to dig down and create some pleasantly surprising results :) and if you have your thinking caps on you can probably see how this stuff would really help to bring a DSL to life for your applications special needs.

And that’s where I’m going to conclude this rambling post… ;o) Hopefully it'll spark some thoughts about what's possible with IronPython and allow you avoid some simple mistakes when using .Net classes in iron python.

Second Alpha

I released the second alpha of the Splicer library last night… you can grab it from the Splicer Codeplex site

Key changes are:

• It’s been FxCop’d – which in turn has seen some naming issues, misspellings, security issues, incorrect disposal implementations etc being fixed.
• Implemented some features:
• Support for adding System.Drawing.Image’s to a track (using overloads of the AddImage(…) methods.
• Support for shadow copying input files – this will take a copy of a clip file, and use that in the ITimeline, disposing of the copy when the clip is disposed – this is especially useful when dealing with source filters which don’t like reading the same file simultaneously.
• AbstractProgressParticipant class which can be used to generate your own progress participants, 0 or more participants can be registered for each group.  This allows you to do things like perform screen captures on the output of the timeline.
• Windows media renderer will now notify you when the selected rendering profile is unsuitable for your timeline.
• Fixed some bugs
• The clock is now correctly disabled for the filter graph when rendering to a file.  This will speed up rendering in many scenarios.
• When supplying an audio encoder for either WAV or AVI output, the format settings for the encoder were not being applied properly (was using defaults, instead of setting the number of channels, khz & kbps).

Though there are also plenty more, I doubt anyone has actually been using this library so far (as it's really only useful in a rather select set of circumstances) - however just to offer some vague hope of encouragement I’ll include some code samples to get you started – as there is no documentation short of the 150 odd unit tests included with the project... and won't be for some time.

Example 1 

First off, lets imagine you have some web 2.0 website, and you encourage people to upload audio comments, instead of text, yet you don’t know what format they’re going to provide, and you want to keep things standard (say by outputting a certain windows media audio format) that eases pressure on your web server.


Here we are creating a Timeline, this is a container for all the audio and video compositing we are doing, next we create a group for our audio, a group is a top level composition… we then add a track to our audio group, which is a container for our audio clip we wish to add (in this case it’s an mp3, but it could be any audio format that you have a codec installed for).

Once your Timeline is ready to be rendered, we create a renderer, in this case it’s a windows media renderer, which takes your timeline, the name of the output file and the windows media profile we wish to use – profiles are xml strings containing the settings for the windows media encoder – A couple of bundled ones are included, though you would probably want to create your own using the “Windows Media Profile Editor” which is included with Windows Media Encoder 9.

Last of all we invoke the Renderer, generally rendering can take a little while, depening on the size of the content, and the complexity of the encoding process - so I would suggest invoking the Renderer asynchronously (all Renderers support the BeginRender(…) … EndRender(…) methods which should make this trivial).  Our examples will use the blocking Render() method, just because it's more concise.

Example 2

So audio cool, what about video?

Much the same:



This time we create two groups, and two tracks, one of each for the audio from our source file, and another of each for the video.

Note we’ve specified the format of the video a little – so it will be 24 bits per pixel, and 320 high by 240 wide.
 
In this case we use a helpful method on the timeline itself which will Add a video clip to the video group and an audio clip to the audio group at the same time, called AddVideoWithAudio… Again we render it to windows media format, but this time we use a profile which has settings for both audio and video.

The bits per pixel becomes important when using something transitions or effects, as you need an alpha channel for them to work - this is generally achieved but just bumping the BPP from 24 to 32.

Example 3

Right, moving on from here, lets cover some more interesting ideas – first off, lets build a slide show video…



In this case we are producing a slideshow – Note that 25 in the constructor of the timline, this is setting the frames per second, in this case we are saying 25fps…  The renderer has the final say in the “rendered” frames per second, based on it’s settings… but some renderers will use the frame rate they receive, such as the windows renderer (which renders the video into a window on-screen)

The slideshow fades between 4 still images, each lasts for a duration of 2 seconds, we add in a couple of transitions, fading out of one clip and into another – last of all we are using an AudioMixer effect on the audio track to create an Audio envelope, which means it will take one second for the audio to fade in from 0.0 volume to 1.0 volume (0 to 100%) and then play until one second before the end of the clip, at which point it will fade out from 100% volume to 0% volume.

Example 4

Last of all, you can capture rendered frames from a video, this might be handy to display on a web page, with links to the various quality formats beneath – you can grab these frames during encoding, or you can render to “null”, which allows you the grab frames, but wont produce any output video or audio file… Lets take a look at the second idea:


Here we add a video group (we don’t need any audio, we’re not capturing sound samples) and then create an ImageToDiskParticipant (this is just an example class, you would probably want to implement your own derived from the AbstractProgressParticipant class).  In this case the format for the video group will be what you end up with, our ImagesToDiskParticipant implements a simple queue where it picks out the nearest frame to the specified times.. so in this case transitions.wmv is an 8 second video clip, and we will take frame "snapshots" at 1,2,3,4,5,6 and 7 seconds aproximately. 

To make use of our participant we use one of the overloads of NullRenderer, and supply it a single-element array containing the image to disk participant.

The end result is 7 jpeg images on disk (frame0.jpg -> frame6.jpg)

Hopefully this has given you some ideas about how Splicer can be used for dealing width audio and video, next time I do an alpha drop I might include some samples which use different container formats (WAV & AVI) - or maybe I'll use splicer as an example of just how not to write an API you wish to expose to IronPython due to it's bloated number of overloads.

hmmm?

edit: seems better now, swapped to a simpler layout though I doub't it will work well for people with lower resolutions... most people will (hopefully) read this via RSS at any rate.

Back once more...

I decided to spend ten minutes implementing a version of the last post which didn't require us to generate a function definition for a python method with a generic number of parameters... it's actually not too hard.

First off, the key to solving this problem is to generate a Type for our custom delegate - for performance reasons you would want to cache the results, but we won't do that here... Also I think it would be more apropriate for the first 0->9 parameters that we just return a generic Func delegate (generic Func & Proc delegates are going to be part of .Net 3.0, as mentioned here - At the moment I use something similar from the RhinoCommons library, have a look here)

So, first off we build a helper method for generating our delegate - just heavy use of emit here, based off an example on MSDN:


Basically it goes through the motions of building a delegate type, derived from MulticastDelegate with our selected output type and parameter types...

Now to put that type to use:


Our test creates our custom delegate type for the number of parameters we have, and then we use that when invoking the PythonEngine.CreateMethod<TDelegate>(string statements, IList<string> parameters) method.

Here we have to do a little reflection, simply because we cant directly use a local variable of type "Type" as a generic type parameter, so we grab the generic method, set the parameter and then invoke it.

However in the end, as you can see we get a real Delegate, which we can use DynamicInvoke upon.

This is actually pretty handy, and it's certainly a lot more script-friendly, compared to the alternative of forcing users to look up values from context by index ie.

    context['user'] + ' is ' + str(context['age']) + ' years old'

vs.

    user + ' is ' + str(age) + ' years old'

I know what I prefer ;o)

Part  2 - More IronPython and delegates…

So we have briefly covererd strongly typed delegates and event handlers... But they make the assumption that we know what our arguments are at compile time, what if we dont?

Say for instance you're expression engine has a bag of contextual values being passed around in a dictionary, like this:


What if we want to use python to evaluate expressions against that context?... Say something like writing out “my name is alex and my age is 25" - the expression in python is easy enough, we can go:

    ‘my name is’ + name + ‘ and my age is ‘ + str(age)

But how do we marry all these together... lets start exploring... ever noticed that delegates have a DynamicInvoke method with a signature like this?

    object DynamicInvoke(params object[] parameters)

Perhaps we can try and use this to our advantage... lets see shall we?

First Attempt

Sadly this doesn’t work so well, seems the PythonEngine is looking for a concrete delegate... we could try giving it void delegate, but it does type checking on the number of parameters and their type, so we're a little stuck.

At this point I think the best answer is to actually build some code which generates the apropriate delegate type, based on the supplied dictionary, and then passing that onto the python engine, but there is another way… it’s a little less elegant, but it’s at least amusing:

Round 2

First of let’s build a little helper method...


You can no doubt see where this is going – so given a statement like this:


We’d end up with a string like this:

    def func(user, age):
        return user + ' is ' + str(age) + ' years old'

And, moving on from there we build a test…