Reading non-integer matrices in python

[eyalbenda]

Hi,

I've been extremely impressed with the work done on BPcells! I wanted to ask if you are planning to adding support for reading and subsetting non-integer matrices into python as well, since that is something that would be super helpful for R/python mixed workflows. Is this currently planned? Are there major hurdles for adding that support?

Cheers and thanks for creating this wonderful package!

[immanuelazn]

Hi @eyalbenda, glad to see you using our package!

The plan for Python is complex, and is on our mid-term to-do list. Essentially the idea is that a lot of our matrix/fragment logic is held on the S3/S4 R OOP implementation, rather than being held in the c++ layer. This means that adding a lot of the logic into Python requires adding duplicate logic.

@bnprks and I had discussed this thoroughly, and decided that we want to do this in a more comprehensive way, rather than just adding in Python capability one function by one.

This means that we are planning on moving tons of the R logic to c++, which will take approximately 1-2 months of working hours. Timeline will likely be for early next year. Once that is done, we can wrap the functions much more easily in the python layer, including any functionality that you are probably interested in using.

Thanks!

Immanuel

[eyalbenda]

Thank you! So just to double check, the logic for reading a float rather than an int matrix would be different beyond what could be solved with templates? the data structures are different and would require a different implementation?