We should allow track ids to be repreated across locations. e.g.
Location 1 has trackID=42
and
Location 2 has trackID=42
This fine and should be expected. As opposed to:
Location 1 has trackID=42
Location 1 has a different cell with trackID=42
This is just an error in the data.
We have some code that does not handle this correctly. I know that datasetLoader.addAggregateColumn at least does not handle this correctly. You can see in the code it is grouping by just trackId.
The way we first noticed this in the UI is with the "Track Length" distribution. Since it groups by just track ID the track lengths (number of cell instances in each track) are way too high. E.g. in an experiment with 72 frames, 72 must be the max length, but we saw a max of 10,000.
We should allow track ids to be repreated across locations. e.g.
This fine and should be expected. As opposed to:
This is just an error in the data.
We have some code that does not handle this correctly. I know that
datasetLoader.addAggregateColumnat least does not handle this correctly. You can see in the code it is grouping by just trackId.The way we first noticed this in the UI is with the "Track Length" distribution. Since it groups by just track ID the track lengths (number of cell instances in each track) are way too high. E.g. in an experiment with 72 frames, 72 must be the max length, but we saw a max of 10,000.