The function to interpolate the ephemerides given spacecraft positions will fail if any extrapolation is necessary. This makes sense to me and it is up to the user to pass in spacecraft ephemeris that spans the range of the light curve.
However, the case for extrapolation also happens without the user's control. In particular, if the light curve is too old, then JPL Horizon may not return ephemerides that spans the light curve.
For example, I pass in a light curve ranging from 2458234.5 to 2459946.5 JD without specifying spacecraft positions. Running a parallax fit, JPL Horizon was downloaded but its timestamps does not span this light curve (see figure).
The function to interpolate the ephemerides given spacecraft positions will fail if any extrapolation is necessary. This makes sense to me and it is up to the user to pass in spacecraft ephemeris that spans the range of the light curve.
However, the case for extrapolation also happens without the user's control. In particular, if the light curve is too old, then JPL Horizon may not return ephemerides that spans the light curve.
The failure happens in line 198 here
https://github.com/ebachelet/pyLIMA/blob/ac9f76ab47d39a64885c80f5fd79327c5a6a180d/pyLIMA/parallax/parallax.py#L183-L204
For example, I pass in a light curve ranging from 2458234.5 to 2459946.5 JD without specifying spacecraft positions. Running a parallax fit, JPL Horizon was downloaded but its timestamps does not span this light curve (see figure).