ToDo's for the for sparse pixels codes

[jonwright]

• Currently there is a single threshold - it makes sense to have a lower threshold for higher angle data.
• Change the compression in the sparse files to be faster for IO
• Fix up the assemble_label code to recycle the previous files (h5py.ExternalLink or VDS)
• Use the new default filenames by default

[jonwright]

• Currently there is a single threshold - it makes sense to have a lower threshold for higher angle data.. There is an option in assemble_label now to have a threshold image

• DONE: Use the new default filenames by default

• clean up the notebooks to use default filenames (add experiment=ma1234, session=yymmdd, and get these from getcwd or silx or blissdata)

• Change the compression in the sparse files to be faster for IO

• Fixing up the assemble_label code to recycle the previous files (h5py.ExternalLink or VDS), this means refactoring the segmentation to be writing into fewer files, or assembling as it goes.

• make pks_table.tocolf and have this be comparable to reading of a columnfile

• move this issue into a projects thing?