angelolab / ark-analysis / 7426

95%

master: 94%

push

travis-ci-com

Clean up Pixie pipeline (#856)

* Remove extraneous pixel_cluster_prefix arg from create_pixel_matrix

* Remove cell_cluster_prefix from notebook

* Make norm_vals_name naming convention consistent with channel_norm prefix

* Change channel_norm.feather format to match channel_norm_post_rowsum.feather

* Rename pixel_norm to pixel_thresh

* Specify weights as SOM weights and remove saving of SOM to meta cluster mapping

* Clarify cell_counts_norm as size normalized in naming conventions

* Save weighted_cell_channel_name as a feather file instead of csv

* Remove extraneous 's' at the end of meta_cluster_count_avgs files

* Rename clustering notebooks to ensure they're identified as Pixie

* Rename notebooks to use up-to-date names

* PYCODESTYLE

* Ensure 'size_norm_name' propagates into notebook

* weights_name to cell_weights_name change in notebook

* Notebook variable adjustments

* Make sure example dataset doesn't have any unnecessary prefixes in it

* Ensure cluster_counts_norm_name passed wherever necessary

* Cell clustering documentation fixes

* Write out cluster counts size norm instead of cell mat for cell cluster notebook test

* Don't write out extraneous cell_mat.feather

* Ensure tests match up with updated sans-cell_mat process (we really need to refactor to OOP...)

* Patch up test_utils dimensions for _make_neighborhood_matri

* Initial pass at reorganizing files

* Remove extraneous print statements

* Reorganize pixie notebooks so example data download is the first thing; renumber pixel clustering notebook

* Ensure example dataset pulls from the correct HuggingFace hash

* Final documentation changes addressed

* Reorganize all notebooks so example dataset download is at the beginning

* Update file names in README.md to reflect updated names in notebooks

* Fix another instance of example_pixel_clustering.ipynb

* Fix more incorrect notebook name paths

* Ton of not... (continued)

63 of 63 new or added lines in 2 files covered. (100.0%)

3312 of 3530 relevant lines covered (93.82%)

0.94 hits per line