`get_2D_distance` errors if provided a track graph but not edge list

[CBroz1]

non_local_detector/src/non_local_detector/analysis/distance2D.py

Lines 197 to 235 in eb86ba0

	def get_2D_distance(
	position1: np.ndarray,
	position2: np.ndarray,
	track_graph: nx.Graph = None,
	edges: list = None,
	precomputed_distance: bool = True,
	) -> np.ndarray:
	"""Distance of two points along the graph of the track.
	Parameters
	----------
	position1 : np.ndarray, shape (n_time, 2)
	position2 : np.ndarray, shape (n_time, 2)
	track_graph : nx.Graph or None
	edges : list or None
	precomputed_distance : bool, optional
	Returns
	-------
	distance : np.ndarray, shape (n_time,)
	"""
	position1 = np.asarray(position1)
	position2 = np.asarray(position2)
	if position1.ndim < 2:
	position1 = position1[np.newaxis]
	if position2.ndim < 2:
	position2 = position2[np.newaxis]
	if track_graph is None:
	distance = np.linalg.norm(position1 - position2, axis=1)
	else:
	node_positions = nx.get_node_attributes(track_graph, "pos")
	node_ids = np.asarray(list(node_positions.keys()))
	node_positions = np.asarray(list(node_positions.values()))
	# remove outer boundary edge
	bin_edges = [e[1:-1] for e in edges]

track_graph and edges args are both optional. If provided the former and not the latter, this results in a TypeError: 'NoneType' is not iterable'. Can edges be inferred from track_graph? Should this instead raise an error about 'if a, must include b'?

[edeno]

It doesn't make sense to specify the edges without the track graph. This could be better documented (like many things in this repo). At some point I will get to the full documentation of everything here.