Yes! I can do it for K2 and I'm guessing it should be similar.
You can see with fits.info(fits_file)
that it has the 84 extensions (21 modules * 4 channels per module) you mention.
Then you can select a channel, e.g. 65, and prepare also the wcs coordinates so you get the RA and dec associated with the image for the axis.
n = 65
with fits.open(fits_file, mode = "readonly") as hdulist:
wcs_info = WCS(hdulist[n].header)
cal_image = hdulist[n].data
header = hdulist[n].header
Then you can just plot:
plt.figure(figsize = (15,15))
fig = plt.figure(figsize=(12, 8))
ax = fig.add_subplot(1, 1, 1, projection=wcs_info)
im = ax.imshow(cal_image, vmin = np.percentile(cal_image,1),vmax = np.percentile(cal_image, 98), origin='lower')
ax.coords[0].set_ticklabel_position('l')
ax.coords[0].grid(color='white', ls='dotted')
ax.set_ylabel('Dec')
ax.coords[1].set_ticklabel_position('b')
ax.coords[1].grid(color='white', ls='dotted')
ax.set_xlabel('RA');