[Numpy-discussion] incremental histogram

[denis]

On 04/05/2010 14:09, Neal Becker wrote:
> denis wrote:

>> Neal,
>>     I like the idea of a faster np.histogram / histogramdd;
>> but it would have to be compatible with numpy and pylab
>> or at least a clear, documented subset (doc first).
>
> The point is not to be faster, it's to be incremental.

OK, different points: I'd like it to be very fast and leverage it

>> Some Wibnis, wouldn't it be nice ifs, for WibniHistogram:
>> - gui with realtime zoom / upsample / smooth: must exist, physicists ?
>> - adaptive binning, e.g. percentiles then uniform
>> - interpolate: fill holes, then *linear or spline

Do any of these make sense / resonate ?

>> += data is nice, but seems orthogonal to histogramming --
>> why not just subclass histogram ?
>>
>
> I thought np histogram was a function, not a class?  To be incremental, it
> has to have state, and so should be a class.

Yes you're right. Is it worth making into a class, with C or Cython ?