CLI scripts #177
CLI scripts #177
Conversation
src/mri/cli/dc_adjoint.py
Outdated
| else: | ||
| log.warn("Trajectory name not found in data header, validation step not done") | ||
| shots, traj_params = traj_reader( | ||
| traj_file, dwell_time=DEFAULT_RASTER_TIME / data_header["oversampling_factor"] |
There was a problem hiding this comment.
you may want to have the raster time being a parameter for future flexibility.
There was a problem hiding this comment.
It already is, will update the code to bring that config here. Good catch
src/mri/cli/dc_adjoint.py
Outdated
| ) | ||
|
|
||
|
|
||
| def recon(obs_file: str, traj_file: str, obs_reader, traj_reader, fourier): |
There was a problem hiding this comment.
Type hints for hydra zen to be happy ?
There was a problem hiding this comment.
That is handeled above. The readers are partially configured objects based on inputs.
src/mri/cli/dc_adjoint.py
Outdated
| per_ch_image = fourier_op.adj_op(kspace_data) | ||
| combined_image = np.linalg.norm(per_ch_image, axis=-1) |
There was a problem hiding this comment.
what if I provided the smaps in the file or separately ?
There was a problem hiding this comment.
Ya, thats not yet supported :P
src/mri/cli/dc_adjoint.py
Outdated
| ) | ||
| per_ch_image = fourier_op.adj_op(kspace_data) | ||
| combined_image = np.linalg.norm(per_ch_image, axis=-1) | ||
| pkl.dump(combined_image, open("dc_adjoint.pkl", "wb")) |
There was a problem hiding this comment.
I would rather have the output format left to the user (.npy, .nii)m, as well as the filename. Once reconstructed, the user will probably want to view the data in their favorite viewer.
Co-authored-by: Pierre-Antoine Comby <pierre-antoine.comby@crans.org>
Co-authored-by: Pierre-Antoine Comby <pierre-antoine.comby@crans.org>
This is an ongoing PR for creating CLI for pysap-mri reconstruct to streamline and make the work reproducibile.