What happened tp scipy.stsci?

[Wanderer]

Back in scipy 0.7 there was a package called stsci that had function
scipy.stsci.image.median that created a median image from a stack of
images. The stsci package is dropped in v0.8. Has this functionality
been moved to a different package?

Thanks

Apologies if this is a double post. I had problems with google groups.

 

[Eelco]

Back in scipy 0.7 there was a package called stsci that had  function
scipy.stsci.image.median that created a median image from a stack of
images. The stsci package is dropped in v0.8. Has this functionality
been moved to a different package?

Thanks

Apologies if this is a double post. I had problems with google groups.

Perhaps that package has been merged with scipy.ndimage?

Either way, if I understand correctly, what you are trying to do could
be done with numpy.median(imagestack, axis=stackaxis), no?

 

[Wanderer]

I found it it is in the stsci package.

Either way, if I understand correctly, what you are trying to do could
be done with numpy.median(imagestack, axis=stackaxis), no?

Yes, I guess so. I didn't realize numpy.median had an axis option.
Thanks. That's one less import.

 

[Wanderer]

I found it it is in the stsci package.



Yes, I guess so. I didn't realize numpy.median had an axis option.
Thanks. That's one less import.

Actually numpy.median doesn't work. numpy.median does not accept axis
greater than 2.

 

[Wanderer]

Actually numpy.median doesn't work. numpy.median does not accept axis
greater than 2.

numpy.median does work. I had to use dstack to create my stacks.

 

[Eelco]

numpy.median does work. I had to use dstack to create my stacks.

Exactly what _stack function to use always confuses me as well . I
prefer to use numpy.concatenate(axis=2), which does exactly the same
as dstack, but is much less confusing to me. Especially useful in
keeping hstack and vstack apart. 'horizontal' and 'vertical', what
does that even mean? One could look it up or try to think about it;
but in my opinion its much simpler to directly deal in terms of axis
numbers.