Marcel,

It sounds like you want use IPython traitlets, something like:

from IPython.utils.traitlets import HasTraits, Float, on_trait_change, Any
from skimage.viewer import ImageViewer

class Pipeline(HasTraits):

       param1 = Float(0.1)
       param2 = Float(0.2)
       param3 = Float(0.3)
       viewer = Any

       @on_trait_change('param1, param2')
       def use_previous_function(self):
            # do something with previous function

       @on_trait_change(''param3')
       def use_next_function(self):
            # do something with next function

v =ImageViewer(...)
p = Pipeline(viewer=v)

On Thursday, January 29, 2015 at 8:35:21 AM UTC-6, Marcel Gutsche wrote:

Hi Juan, 

thanks for the hint. That could work, but i forgot to mention, that the previous function is quite expensive to compute, so I would like to update the first one only, if the slider for its parameter is touched. Right now i try to reduce the amount of recomputations with memoization using the joblib module, but I think I need a while to figure out when exactly values are cached.

Marcel  

Am Donnerstag, 29. Januar 2015 12:01:14 UTC+1 schrieb Juan Nunez-Iglesias:
Hi Marcel,

I'm guessing you only need to see the final output of the pipeline? Why not define your own function?

def piped(image, parameter1, parameter, another_parameter):
    a, b = previous_function(image, parameter1, parameter)
    new_image = next_function(image, a, b, another_parameter)
    return new_image

Then use that in your plugin?

Juan.




On Thu, Jan 29, 2015 at 7:58 PM, Marcel Gutsche <marcel....@gmail.com> wrote:

Hi folks,

kind of a follow up question of my last one (see here). I still have a chain of image processing tasks, which I want to visualize with the image viewer. Certain parameters should be changed on the fly, where i can see the results in the image viwer immediately by the use of plugins. Something like this basically:


a, b  = previous_function(image, parameter1 = 0.002
                                    parameter = 0.02)
new_image = next_function(image, a, b, another_parameter = 3)


I managed to pass the values a and b, generated by a previous function using Juan's hint with functools.

next_function

which was fine as long as I don't want to change parameter1 and parameter2 interactively. What do I do now, when I want to change them? 

Using functools here won't get me far, since now I want to be able to change parameter1 and parameter2 with a slider, as well as another_parameter.

Maybe there is still a way to achive this, I saw something like a callback argument for the widgets, but I am not sure how to use them or whether they are related to my problem at all. I would also be fine with writing my own Plugin class, but could not get my head around the given examples. I feel that something like this is quite common, so I hope there are already ways to achieve this.

Thanks a lot for the great work,
Marcel

--
You received this message because you are subscribed to the Google Groups "scikit-image" group.
To unsubscribe from this group and stop receiving emails from it, send an email to scikit-image...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

​