Interval overlap (signed distance) for numpy lists of intervals

overlap.py

import numpy as np
from numpy import typing as npt
def overlap(a:npt.NDArray, b:npt.NDArray, x:npt.NDArray, y:npt.NDArray):
   """Compute the signed distance between lists of intervals"""
    overlap_min = np.maximum(a, x.reshape(-1,1))
    overlap_max = np.minimum(b, y.reshape(-1,1))
    signed_overlap_len = overlap_max - overlap_min
    return signed_overlap_len
  

ab = np.array([
    [100,200], # <-- represents interval from 100 to 200
    [200,300],
    [300,400]
])

xy = np.array([
    [50,120],
    [120,310],
    [350,450],
    [450,550]
])

overlap(ab[:,0],ab[:,1],xy[:,0],xy[:,1])
# array([[  20,  -80, -180],
#        [  80,  100,   10],
#        [-150,  -50,   50],
#        [-250, -150,  -50]])

# Note 
# - positive values in the output are overlap distance
# - negative values in the output are minimum separation distance
# - the columns in the output correspond to the rows in `ab`, and 
# - the rows    in the output correspond to the rows in `xy`
# - zero values indicate either
#   - touching intervals, OR
#   - one of the inputs was zero length
# - if the input contains intervals which are
#   - zero length; the outputs will be AMBIGUOUS
#   - reversed;    the outputs will be INVALID