`cube.get_tpf()`: `IndexError` when `frame_range` is specified and `calculate_poscorr=False`

[orionlee]

An IndexError is raised from cube.get_tpf() when frame_range is specified and calculate_poscorr=False

Workaround: setcalculate_poscorr=True

---

Example:

%%time

from astropy.coordinates.sky_coordinate import SkyCoord
import astropy.units as u
import numpy as np

from tesscube import TESSCube

# TIC 400621146, sector 45
ra, dec = 95.954934, 31.679387
sector, camera, ccd = 45, 1, 2

coord = SkyCoord(ra * u.deg, dec *u.deg, frame="icrs")
cube = TESSCube(sector=sector, camera=camera, ccd=ccd)

# get only 1 good quality frame
frame_day3 = np.where((cube.time > cube.time[0] + 3) & (cube.quality == 0))[0][0]
hdul = cube.get_tpf(coord, shape=(11, 11), frame_range=(frame_day3, frame_day3 + 1), calculate_poscorr=False)
hdul

The error and the stack trace.

C:\dev\tesscube\src\tesscube\query.py:52: RuntimeWarning: overflow encountered in scalar multiply
  pixel_offset = (column - 1) * self.nframes * self.nsets + (
---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
File <timed exec>:18

File C:\dev\tesscube\src\tesscube\cube.py:296, in TESSCube.get_tpf(self, target, shape, frame_range, frame_bin, calculate_poscorr)
    288 else:
    289     k = np.ones(len(self), bool)
    291 time, timecorr, cadenceno, quality, pos_corr1, pos_corr2 = (
    292     self.time[k],
    293     self.timecorr[k],
    294     self.cadence_number[k],
    295     self.quality[k],
--> 296     pos_corr1[k],
    297     pos_corr2[k],
    298 )
    300 if frame_bin > 1:
    301     idxs = np.hstack(
    302         [
    303             r[: frame_bin * np.floor(len(r) / frame_bin).astype(int)]
   (...)
    307         ]
    308     )

IndexError: boolean index did not match indexed array along dimension 0; dimension is 1 but corresponding boolean dimension is 3450

---

Version: 1.0.2dev0
Platform: Windows 11