Closed lpawela closed 4 months ago
That should be doable, but would need you to change the way you format the data. If the matrix is 3 separated datasets: I, J, and V; like the above example, we have to combine them somehow to visualize them. This is impossible for the current NiceGUI framework (i.e. these actions: select I, select J, and select V is not doable with the current framework). If you can change matrix_coo = tuple(I,J,V) or matrix_coo = np.asarray([I, J, V]), I can add the requested visualization capability to broh5
Let reopen this when it's doable.
It would be nice to be able to visualize sparse matrices, especially in the COO format. For instance if we have the following structure: :card_index_dividers: HDF5.File: (read-only) ├─ :open_file_folder: some_group │ ├─ :open_file_folder: matrix_coo │ │ ├─ :1234: I │ │ ├─ :1234: J │ │ └─ :1234: V
Also, support for matrices indexed from 1 would nice for this case.