Weights argument error when using multifit

[Andrew Thron]

Francisco, 

Sorry, I found one other bug. When I use the multifit routine to fit a Voigt function to my ZLP I keep getting the error message "DeprecationWarning: The `weights` argument is deprecated and will be removed in the next release". The error message appears continuously until the fitting routine stops. I am not using any weights for the fitting routine. I was using the MPFIT optimizer with bounds, and I tried removing the bounds and switching to the least squares optimizer, but I still get the same message. I also specified weights=None for the multifit routine, but the error message still shows up. The multifit routine still fits Voigt function well, I just get hundreds of lines of text from the error message showing up on the notebook. 

Cheers, 

Andrew 

[Francisco de la Peña]

Andrew,

Thank you for reporting this. I have tried to reproduce it but I couldn't. Which version of HyperSpy are you using? Could you send us the code that triggers the bug?

Thanks!

Francisco

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

[Andrew Thron]

Francisco, 

I am using the latest version of HyperSpy 0.7 downloaded from the link on the website. I downloaded the Python installer, rather then the "Full installer" , and I installed it with my Anaconda software package. The code that trigger the bug is in the model  module in Hyperspy. I attached module to the post. 

Cheers, 

Andrew 

[Andrew Thron]

Francisco, 

One last detail. The error message occurs on line 977 of model module. 

Cheers, 

Andrew 

On Wednesday, April 16, 2014 1:47:32 AM UTC-7, Francisco wrote:

[Tomas Ostasevicius]

Hi Andrew,
Thanks for the bug report!
You get the weights error because it was (apparently) not removed cleanly enough.
I implemented a quick fix in https://github.com/hyperspy/hyperspy/pull/281
so as soon as it is merged, the latest version should be without the bug

Best wishes,
Tomas

[Francisco de la Peña]

Hi Andrew and Tomas,

I have just merged Tomas' fix. As usual you can upgrade to the development version using the following command:

    pip install --upgrade https://github.com/hyperspy/hyperspy/archive/master.zip

Thank you Andrew for reporting this and Tomas for fixing it,

Francisco