Create a keogram for 24h of images

keogram.py

#!/usr/bin/env python3
import sys
import tqdm
from pathlib import Path

import numpy as np
import matplotlib.pyplot as plt

import astropy.units as u
from astropy.time import Time

if __name__ == "__main__":
    p = Path(".")

    date = sys.argv[1]

    fpaths = list(p.glob(f"../**/{date}T??:??:??.???.jpg"))

    tstr = [fpath.name.replace(".jpg", "") for fpath in fpaths]
    t = Time(tstr, format="isot", scale="utc")
    idx = np.argsort(t.mjd)
    
    nfiles = len(t)
    for i in tqdm.tqdm(range(nfiles)):
        fname = fpaths[idx[i]]
        try:
            img = plt.imread(fname)
            ny, nx, nz = img.shape
            if i == 0:
                raw_keogram = np.zeros(ny * nfiles * nz).reshape(ny, nfiles, nz).astype(img.dtype)
            raw_keogram[:, i, :] = img[:, nx // 2, :]
        except:
            print(f"Error processing {fname}")
            pass

    plt.imsave(f"raw_{date}.png", raw_keogram)

    tplots = Time(f"{date}T00:00:30", format="isot", scale="utc") + np.arange(1440) * u.min

    keogram = np.zeros(ny * 1440 * nz).reshape(ny, 1440, nz).astype(img.dtype)
    for i, tplot in enumerate(tqdm.tqdm(tplots)):
        dt = t[idx] - tplot
        c = np.abs(dt.to(u.s).value) < 30

        if np.sum(c) > 0:
            mean = np.mean(raw_keogram[:, c, :], axis=1)
            keogram[:, i, :] = mean

    plt.imsave(f"keogram_{date}.png", keogram)