Big feature spun off of a comment made by @toloudis! This is part of the changes in #26.
Expected review time: large, 30-40 min
This allows the writer to take in an INDETERMINATE feature type, and attempt to infer the feature's type from the data values.
It also adds an extra validation step to the writer, where it checks the data type before writing. Features will be skipped if they have incorrect types, based on requested behavior by scientists.
I recommend starting with writer.py to see the high-level view of the changes that happened, and then looking at utils.py for the new utility methods. There are also unit tests added in test_utils.py.
Big feature spun off of a comment made by @toloudis! This is part of the changes in #26.
Expected review time: large, 30-40 min
This allows the writer to take in an
INDETERMINATE
feature type, and attempt to infer the feature's type from the data values. It also adds an extra validation step to the writer, where it checks the data type before writing. Features will be skipped if they have incorrect types, based on requested behavior by scientists.I recommend starting with
writer.py
to see the high-level view of the changes that happened, and then looking atutils.py
for the new utility methods. There are also unit tests added intest_utils.py
.I've tested these new changes with Derek's dataset; you can view the first few datasets converted here: https://dev-aics-dtp-001.int.allencell.org/nucmorph-colorizer/dist/index.html?collection=https%3A%2F%2Fdev-aics-dtp-001.int.allencell.org%2Fmicroscopy%2FClusterOutput%2FH2B_Deliverable_AnalysisPipelineOutput%2FH2B_Deliverable_InputImages_123Timelapses_Composite_Output%2Fcolorizer%2Fcollection.json&dataset=3500005807_1&feature=Col4MigratoryCell&color=matplotlib-cool&palette-key=adobe