Frame Utilities

This page documents the frame assembly helpers from nbs/02_frame_utils.py.

Normalize RAW Axis Order

import numpy as np

from mertisreader.frame_utils import normalize_raw_axis_order, get_frame_dimensions, detect_interpolation_target, assemble_frames

raw_arr = np.arange(24, dtype=np.float64).reshape(2, 3, 4)
normalized = normalize_raw_axis_order(raw_arr)

print(raw_arr.shape)
print(normalized.shape)

(2, 3, 4)
(4, 3, 2)

Describe Frame Dimensions

frames_dict = {
    "sample_a": np.ones((2, 3, 4), dtype=np.float64),
    "sample_b": np.zeros((2, 3, 2), dtype=np.float64),
}

dims = get_frame_dimensions(frames_dict)
target = detect_interpolation_target(frames_dict, mode="up")

print(dims)
print(target)

          n_files  n_pixels  n_spectrals
sample_a        4         3            2
sample_b        2         3            2
{'n_spectrals': 2, 'n_pixels': 3, 'mode_applied': 'up', 'reason': 'Upsampling to largest frame'}

Assemble Frames

cube, metadata = assemble_frames(frames_dict, interp_mode="up")

print(cube.shape)
print(metadata)

(2, 3, 6)
{'interpolation_mode': 'up', 'interpolation_order': 1, 'target_n_spectrals': 2, 'target_n_pixels': 3, 'n_temporal_samples': 6, 'reason': 'Upsampling to largest frame', 'file_keys': ['sample_a', 'sample_b'], 'file_indices': [(0, 4), (4, 6)]}

Notes

The module keeps the axis handling explicit so RAW, CAL, and PAR data can be normalized and assembled with the same downstream workflow.