To-do list: revise and review MACE and Dask interface.

[dyang37]

• Review and revise MACE 3D interface without dask functionality.

  1. Revise the function signature in the way that it matches up with the recon() function signature style, with proper line breaks and same type of arguments grouped together.
  2. Think about how the default value of image_range should be decided.
  3. Revise doc string.
  4. think about how to define denoiser() subroutine that accepts (user defined) denoiser function as an input.

• Work on 3D MACE demo to achieve the following goal:

  1. Increase the contrast of the phantom image so that it is sharp.
  2. Default parameter values should result in good reconstruction images. We modified the constant of auto_sigma_p from 1.0 to 2.0, we need to test it to make sure the reconstruction looks good.
  3. Interface should be well defined.
  4. There should be no hardcoding anywhere. Use config files to specify: download url, file/data path, geometry/sinogram/reconstruction parameters, etc.

• Review and revise python dask interface.

• Design MACE 4D interface based on MACE 3D and python dask interface.

• Design and implement multi-node functionality inside MACE 4D.

[dyang37]

This issue is moved to four new issues: Review and revise MACE 3D interface: #26 Develop MACE 3D demo: #27 Review and revise python dask interface: #28 Design and develop multi-node MACE 4D functionality: #29