apply-rotation-matrix

Apply a rotation-matrix stack to a vector series

generate_fac_matrix, sliding-window analyze_minvar_coordinates, and model-based LMN all emit an (N,3,3) rotation-matrix stack — but until now there was no way to apply it to a field/velocity vector within the toolset. This skill closes that loop: matrix stack + vector series in → rotated vector series out (FAC, LMN, boundary-normal, or any custom frame).

When to use

• "Rotate B (or V) into field-aligned coordinates using the FAC matrix from generate_fac_matrix."
• "Apply my MVA / LMN rotation to the field time-series."
• Any "I have a rotation matrix and a vector, give me the rotated vector" step.

Tool chain (all existing)

(generate_fac_matrix | analyze_minvar_coordinates | model-LMN) → matrix .npy/.npz

• a vector CSV (e.g. from fetch_data_product) → small tvector_rotate call → rotated CSV → render_tplot.

Backend (verified I/O contract)

pyspedas.tvector_rotate(mat_var_in, vec_var_in, newname=None):

• Inputs: two tplot variables — mat_var_in an (N,3,3) matrix stack, vec_var_in an (M,3) vector series. The matrix stack is automatically interpolated (qslerp — quaternion SLERP) onto the vector's timestamps, so the two need not share a time grid.
• Returns: a list of new tplot variable name(s) (e.g. ['vec_rot']) — NOT the array. Retrieve the rotated vectors with get_data(name).
• Designed for fac_matrix_make/generate_fac_matrix output, but works for any rotation stack (MVA, LMN, model).

Procedure

1. Have a matrix stack + a vector series. The matrix comes from generate_fac_matrix (FAC, (N,3,3) .npy), or sliding-window MVA, or a model-LMN step. The vector is a fetched field/velocity CSV (time + 3 components).

2. Load both as tplot variables (small local call): store_data('mat', data={'x':mat_times,'y':mat_NxN3x3}) and store_data('vec', data={'x':vec_time,'y':vec_M x3}). If the matrix has no explicit time axis, give it the vector's time grid (no interpolation needed); otherwise let qslerp handle the mismatch.