Refactor to use AnnData
#56
Labels
enhancement
New feature or request
medium difficulty
a bit more challenging than "good first issue"
This Pleases The John
Stuff John really wants (usually idiomatic refactoring) but isn't critical
Not sure if when we'll do this, but had a thought going to bed last night that we could simplify the model input and clean up the code by decoupling things like factor specification from the inputs (i.e. why are factors hardcoded ? What if someone else wants to use different factors without needing to use the python package directly and override factors when calling the model? What if they have different data? Etc)
We'd do this by converting the input into an Anndata object:
https://anndata.readthedocs.io/en/latest/
X - Matrix contains numeric data
obs matrix contains info like Time, State for each observation
var matrix contains information about variables, such as alternate name, factor assignment, etc.)
So the input could be generalized to that input (or for general users), 2-3 Excel files, one for each matrix.
The text was updated successfully, but these errors were encountered: