3c0483bd26e0bb7654472c514c0b2106906ca050,features/eolearn/features/doubly_logistic_approximation.py,DoublyLogisticApproximationTask,execute,#DoublyLogisticApproximationTask#Any#,57

Before Change

            valid_parameters.append((times[valid_data_mask[:, ih, iw].squeeze()],
                                     data[:, ih, iw][valid_data_mask[:, ih, iw].squeeze()]))

        pool = multiprocessing.Pool(processes=multiprocessing.cpu_count())
        all_parameters = pool.map(self._pool_wrapper, valid_parameters)

        all_parameters = np.reshape(all_parameters, (h, w, 7))

        eopatch.data_timeless[self.new_feature] = all_parameters

After Change

        all_parameters = np.zeros((height, width, 7))

        for height_ind, width_ind in it.product(range(height), range(width)):
            valid_curve = data[:, height_ind, width_ind][valid_data_mask[:, height_ind, width_ind].squeeze()]
            valid_times = times[valid_data_mask[:, height_ind, width_ind].squeeze()]

            all_parameters[height_ind, width_ind] = self._fit_optimize(valid_times, valid_curve)

        eopatch.data_timeless[self.new_feature] = all_parameters

        return eopatch

Instances