You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Probably use the same preprocessing steps as I did for the latest TAIS data (in Andrew's OneDrive but probably shared somewhere on Teams), which should generalize to all HALO input files.
The text was updated successfully, but these errors were encountered:
Figure out what about the preprocessing (it's likely the large size of the total dataset) turns integer columns into floats, e.g., thresholded marker columns. To address that probably just cast back to the original dtypes after the concatenation is complete.
This could be an entire new app in the suite that allows the user to check that the required columns are there, to select the columns to import, and/or perform exploration on the dataset that is to be imported, showing a visual output of the dataset or at least of a sample.
Relatedly, this would allow dataset_formats.py to always work and never die if something is unexpected and would then fail. Otherwise, we need to have the app gracefully die if dataset_formats.py fails.
Probably use the same preprocessing steps as I did for the latest TAIS data (in Andrew's OneDrive but probably shared somewhere on Teams), which should generalize to all HALO input files.
The text was updated successfully, but these errors were encountered: