Adding support for Spacewalk files in tadbit model

[iagomaceda]

Hi,

It would be nice to add support for the Spacewalk binary files as an output for 'tadbit model'.

I already coded the function needed using the ensembl of models as a base:

def write_spacewalk(model_ensemble : StructuralModels, sw_filename : str):
    """
    This function writes a spacewalk file from a model ensemble.

    Parameters
    ----------
    model_ensemble : StructuralModels
        A model ensemble object.
    sw_filename : str
        The name of the spacewalk file to be written.
    """

    region_chr = model_ensemble.description["chromosome"]
    region_start = model_ensemble.description["start"]
    region_reso = model_ensemble.description["resolution"]

    metadata = {"format" : "sw1",
                'version' : '1.0.0', 
                'author' : 'tadbit',
                'date' : str(datetime.now()),
                "name" : f"{model_ensemble.description['project']}_{model_ensemble.description['identifier']}",
                "genome" : str(model_ensemble.description['assembly'])}

    with h5py.File(sw_filename, 'w') as swbf:

        header_group = swbf.create_group('Header')
        header_group.attrs.update(metadata)
        header_group.attrs['point_type'] = 'single_point'

        root = swbf.create_group(metadata['name'])
        genomic_position_group = root.create_group('genomic_position')
        spatial_position_group = root.create_group('spatial_position')

        region_list = [[region_chr, str(region_start + (i - 1) * region_reso), str(region_start + i * region_reso)] for i in range(0, model_ensemble.nloci + 1)]

        genomic_position_group.create_dataset('regions', data=region_list)

        live_contact_map_vertices = np.empty((0, 3), dtype=float)
        
        for model_index, model in enumerate(model_ensemble, start=0):
            
            xyz = np.array([model['x'], model['y'], model['z']], dtype=float).T
            live_contact_map_vertices = np.vstack((live_contact_map_vertices, xyz))

            dataset_name = f't_{model_index}'
            spatial_position_group.create_dataset(dataset_name, data=xyz)

        root.create_dataset('live_contact_map_vertices', data=live_contact_map_vertices)

If you could add it (and the argument to get it) it would be very nice. Feel free to change variable names to fit the rest of the code.

Thanks,

Iago

[zozo123]

Great idea!

Here's the binary format specification: The Spacewalk Binary File Format