Scalar factor in thermal gain drift implementation

[anand-avinash]

The Tdrift here and dT few lines below should be arrays. https://github.com/hpc4cmb/toast/blob/d208ea83fbb09d0b0f31622840e764bb32cd5492/src/toast/ops/sim_gaindrifts.py#L218 But with using thermal_fluct[mask][0], we get only scalars. It passes the test as it is still compatible to multiply with numpy arrays.

[anand-avinash]

Turned out that thermal_fluct[mask] gives a nested array, so using its first index is desirable.