[Bf-committers] Numpy integration

[Campbell Barton]

our existing pixel access is intentionally slow & safe (since blender
could free the image in a thread if we expose the memory directly).

so for numpy access I'm not sure its possible to do in a safe way
(numpy does support some IIRC locking but Im not sure on details)...

to answer your question - yes, it is possible...

suggest bpy_types.py would have a something like this...

|class Image(ID):
|    @property
|    def pixels_fast(self):
|    try: import numpy
|    except: ...fail...
|    return tricky_numpy_wrapper(self)

alternately you can use ctypes to find the pointer and bundle with a
script - since officially this wont be accepted *(dodgy but
functional).

as for pypy, they still dont do py3 or have a CAPI so we are a way off
being able to use it for now.

On Tue, Apr 24, 2012 at 8:52 AM, Alex Fraser <adfries at vpac.org> wrote:
> ----- Original Message -----
>> From: "Campbell Barton" <ideasman42 at gmail.com>
>>
>> clarifications...
>> - numpy would be a python module available but not loaded on startup
>> [...]
>> drawbacks
>> - numpy took a long time to upgrade to python3 (~3 years after py3
>> release), relying on numpy could be a drawback from upgrading python
>> in the future, or moving to a different implementation (like pypy).
>
> I'm mostly interested in it as a way to speed up access to pixel data. To get that, wouldn't we need to have a stronger dependency on it, so that Image.pixels can return a numpy array? Also, given my interest in speed, I'd like to keep the option to move to pypy open. Perhaps that conflicts with my first point ;)
>
> Cheers,
> Alex
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers



-- 
- Campbell