Weird duplication

[ccarouge]

The following code repeats itself:

frontdetection/fronts.py

Lines 151 to 190 in 250f598

	while insearchdist:
	mindist=np.inf
	for jj in range(ptcount):
	if not_used[jj]:
	dist=sp_dist.euclidean((inpts[t]),(inpts[jj]))
	if dist>0 and dist<mindist:
	mindist=dist
	rec=jj
	distr=dist
	#have found nearest unused point
	if mindist<searchdist:
	not_used[rec]=False
	templat.append(inpts[rec,0])
	templon.append(inpts[rec,1])
	t=rec
	else:
	insearchdist=False
	# search other direction
	t=ii
	insearchdist=True
	while insearchdist:
	mindist=np.inf
	for jj in range(ptcount):
	if not_used[jj]:
	dist=sp_dist.euclidean((inpts[t]),(inpts[jj]))
	if dist>0 and dist<mindist:
	mindist=dist
	rec=jj
	distr=dist
	#have found nearest unused point
	if mindist<searchdist:
	not_used[rec]=False
	templat2.append(inpts[rec,0])
	templon2.append(inpts[rec,1])
	t=rec
	else:
	insearchdist=False
	if len(templat2)>0:
	templat=templat2[::-1]+templat
	templon=templon2[::-1]+templon

We go twice through the list of points. Each time the code identifies all the points that are close enough to the point of interest and sort them from closest to farthest. I am not sure the second repetition will find anything since the first search seems to remove all the points that are close enough. The comment says "in the other direction" but the code doesn't seem to give a specific direction to the search.

There are probably ways to rewrite this more efficiently to get the same result as current. But we may need some more information on what the code is supposed to be doing afterwards to ensure the code is correct.

[ScottWales]

From what I understand, the problem is that you're not necessarily starting at the start of a line, just some random point on it. So going only one direction will not find the full line, you have to go back to the start location and try searching the other direction as well.