Improving performance of shapely crosses

Question

I am unsure if this is the fastest way to test if LineStrings cross each other. Is there a better method.

for line_1, line_2, in itertools.combinations(lines, 2):
    crosses = line_1.crosses(line_2)
    if crosses is True:
         print(line_1, line_2)

Obviously this gets pretty slow with large files, I have found maybe loading it into an rtree index may speed this up.

Answer 1

I have improved this now using shapely prep

import shapely
line_lo = [line for line in lines]   
for line in lines:
    line_lo.remove(line)
    line_prep = shapely.prepared.prep(line)
    hits = list(filter(line_prep.intersects, line_lo))
    if len(hits) > 0:
          for line1, line2 in combinations(hits, 2):
              crosses = line_1.crosses(line_2)
              if crosses is True:
                  print(line_1, line_2)

This has significantly improved the processing time on larger datasets, small increases in processing time on small datasets.