Design code structure in repo

unitaryfund / aquapointer

Package applying quantum algorithms to find locations of water molecules in a protein cavity

GNU General Public License v3.0

8 stars 1 forks source link

Design code structure in repo #2

Closed nathanshammah closed 10 months ago

nathanshammah commented 11 months ago

Tidy up code repo so that things are nicely ordered. Aquapointer's current structure, high level looks like this:

MUP1 RISM Slices and points from
Aquapointer
- density_mapping
- processor
- qubo_solution
- water_placement
Notebooks Currently mostly the digital QUBO tests with VQE and QAOA
Registers
Src
- qubo
- qubo_utils
- vqe
- load_data

Let's discuss here whether there are:

modules/folders that could be moved/merged
higher-level divisions to make, e.g., digital and analog
rules on what goes where, e.g., src, notebooks, utils, etc.

Happy to help here. Nice work already!

nathanshammah commented 11 months ago

Here is a first impression on a possible redesign, quite tentative:

aquapointer
- data
- MUP1
- analog
- density_mapping
- processor
- qubo_solution
- water_placement
- analog_digital
- digital
- qubo
- utils
  - load_data
  - qubo_utils
  - vqe
- notebooks
- registers

I am not copying mundane files and folders like tests and init. @Misty-W @FarLab, curious to discuss and get your input.

nathanshammah commented 10 months ago

@FarLab @Misty-W let me know what do you think with this proposed structure and I can open a PR if it looks fine.

FarLab commented 10 months ago

Hi @nathanshammah , I agree with this structure! Just not sure what analog_digital directory is for. I don't think we have an algorithm that fits that description right?

Misty-W commented 10 months ago

Just not sure what analog_digital directory is for. I don't think we have an algorithm that fits that description right?

@FarLab I would use analog_digital for shared files and functions. @nathanshammah do you agree or did you have something else in mind?

FarLab commented 10 months ago

Oh I see, it is meant for files and functions that both the digital and analog algorithm uses. That makes sense!

Misty-W commented 10 months ago

@nathanshammah assuming analog_digital is for common files, I would place processor inside because it contains both digital and analog processors.

Otherwise it looks fine.

nathanshammah commented 10 months ago

@Misty-W I am actually not sure, I copied Will's slide as I thought in the future we may have digital-analog stuff, à la Qadence (I guess?).

Misty-W commented 10 months ago

The possibility of hybrid digital-analog did come up at one point, @FarLab do you remember any suggestions along those lines?