tkimhofer / pym8 Goto Github PK

Hi T,

The random lines remain an issue even with no pre-processing

:

Hope that helps

1. Needs get_idx updated for reading/utility.
2. Import re
3. If you wanted- a comment on the dimension required for X2 when plotting :)

Hi T,

The PCA (analyse.pca) had issues running with ppm values below zero.

This was the code that worked-

idx=np.where(np.std(X, 0)!=0)[0]
X1=X[:,idx]
ppx=ppm[idx]

:)

Hi Torben,

Please note the following-

1. When you pre-define the chemical shift the visualizations appear like this-
   
   I am unsure if you would like them to appear across the full image length.

2. PCA scores plot does not have the option to label data points/customise colour groups.

Many thanks,
C

Hi T,

I only got this function to work after-

1. [0] removed from get_idx
2. mm8.utility added to baseline
3. mm8.utility added to "tf.cast(doub...)

Works fine now I believe.

Also can't get Marias PCA's to run on her computer (they run on mine)

• error is with sklearn/validation
• "ValueError: input contains NaN, infinity or a value too large for dtype('float64')