compress_shape.py
822297d0cbf0master

compress_shape.py
View Options

			import numpy as np

			from pulseq.core.holder import Holder


			def compress_shape(decompressed_shape):
			"""
			Returns a run-length encoded compressed shape.

			Parameters
			----------
			decompressed_shape : ndarray
			Decompressed shape.

			Returns
			-------
			compressed_shape : Holder
			A Holder object containing the shape of the compressed shape ndarray and the compressed shape ndarray itself.
			"""

			if decompressed_shape.shape[0] != 1:
			raise ValueError("input should be of shape (1,x)")
			if not isinstance(decompressed_shape, np.ndarray):
			raise TypeError("input should be of type numpy.ndarray")

			# data = np.array([decompressed_shape[0][0]])
			# data = np.concatenate((data, np.diff(decompressed_shape[0])))
			data = np.hstack((decompressed_shape[0][0], np.diff(decompressed_shape[0])))

			# TODO
			mask_changes = np.hstack((1, np.abs(np.diff(data)) > 1e-8))
			vals = data[np.nonzero(mask_changes)].astype(float)
			k = np.array(np.nonzero(np.append(mask_changes, 1)))
			k = k.reshape((1, k.shape[1]))
			n = np.diff(k)[0]

			n_extra = (n - 2).astype(float)
			vals2 = np.copy(vals)
			vals2[np.where(n_extra < 0)] = np.NAN
			n_extra[np.where(n_extra < 0)] = np.NAN
			v = np.array([vals, vals2, n_extra])
			v = np.concatenate(np.hsplit(v, v.shape[1]))
			finite_vals = np.isfinite(v)
			v = v[finite_vals]
			v_abs = abs(v)
			smallest_indices = np.where(v_abs < 1e-10)
			v[smallest_indices] = 0

			compressed_shape = Holder()
			compressed_shape.num_samples = decompressed_shape.shape[1]
			compressed_shape.data = v.reshape((1, -1))
			return compressed_shape