BF Add ordinal prefix to tmp files and sort appropriately to maintain…

[kastman]

Hi Anisha,

Two small fixes here - the first is just removing some diffs that were somehow committed, causing the example.ipynb to break and be invalid json.

The second is a trick to force tmp files to always be sorted consistently. This allows you to use lists of both images and options that will match in the same order. What was happening before was the the images were being copied into a local dir with a unique, autogenerated names. But this actually changed the order of the images stochastically, so that the keys were mapping back haphazardly, causing each LUT option to map to each image randomly. To fix this I appended a prefix to each one and sort them, so that the first image in the list always has the first tmp image, etc., and the key mapping works correctly. This allows you to match options and images, e.g. for positive and negative maps:

image_options_hot_cold = [
    {'lut': 'Greyscale'}, 
    {'lut':'Red Overlay'}, 
    {'lut': 'Blue Overlay'}]
Brain([atlas, 'zstat1_pos.nii.gz', 'zstat1_neg.nii.gz'], 
      image_options=image_options_hot_cold)

Otherwise, the tmp filenames map all weird, and re-running will shuffle the three option lists so that they map randomly at each call to the notebook cell.

[akeshavan]

@kastman thanks for the PR!

[kastman]

Thanks for putting nbpapaya together! It's helping me finally tell a good story about old data I've been avoiding. I'll send any more PRs that I might come across, but that was the big one so far...

…

On Thu, Jun 15, 2017 at 3:09 PM, Anisha Keshavan ***@***.***> wrote: @kastman <https://github.com/kastman> thanks for the PR! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#9 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACvKxDUpi9YalkE0YfqFAeJ6a8Np3MRks5sEYGEgaJpZM4N6Zjf> .