«

I've been walking up, down, and around instances in the object model
of PIL trying to figure out how to easily calculate how many unique
colors are used in an image, specifically a bitmap (mode "P"). Am I
missing something?

Thanks python friends!

(Larry) wrote in message news:<. com>...
> I've been walking up, down, and around instances in the object model
> of PIL trying to figure out how to easily calculate how many unique
> colors are used in an image, specifically a bitmap (mode "P"). Am I
> missing something?
>
> Thanks python friends!

So far I have: colors=len(sets.Set(list(im.getdata())))

That makes me want to throw up.

At some point, (Larry) wrote:

> (Larry) wrote in message news:<. com>...
>> I've been walking up, down, and around instances in the object model
>> of PIL trying to figure out how to easily calculate how many unique
>> colors are used in an image, specifically a bitmap (mode "P"). Am I
>> missing something?
>>
>> Thanks python friends!
>
> So far I have: colors=len(sets.Set(list(im.getdata())))
>
> That makes me want to throw up.

Probably very inefficient. Have a look at im.histogram(). Something
like:

colours = sum( [ 1 for h in im.histogram() if h != 0 ] )

--
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke
|cookedm(at)physics(dot)mcmaster(dot)ca

"Larry" <> wrote:

> I've been walking up, down, and around instances in the object model
> of PIL trying to figure out how to easily calculate how many unique
> colors are used in an image, specifically a bitmap (mode "P"). Am I
> missing something?

colors = len(filter(None, im.histogram()))

</F>