Matthew Harrigan
Matthew Harrigan
I also want to start splitting up `featurizer.py`--which is huge--into files grouped by theme
This just splits out the rmsd ones. We can do it bit by bit
Realized I'm going to have to do this in libdistance so it works with clustering
+0 With all the plotting options the command line might get unwieldy
I don't think we should have two codepaths (2) for maintainability. How often do people have labels that aren't integers? I feel like the huge majority of this code goes...
Hi, We (the developers) typically work on mac os or linux, which have great support for symlinks. You may run into problems running the tutorial on windows. You can (1)...
You'll probably have to modify each of the generated scripts in a similar way
``` python sklearn/pipeline.py in _transform_one(transformer, name, weight, X) ``` vs ``` python delayed(sklearn.pipeline._transform_one)(trans, name, traj_list, self.transformer_weights) ``` @cxhernandez told me this in person, but now I can see it clearly...
cc #733
Runtime dependencies only available on omnia: - ~~fastcluster~~ - pyhmc