Open ap-ack opened 3 months ago
Are you able to provide some timing comparisons, to see how much performance is impacted by this change? With the way you did it, I'm guessing it's minimal.
I measured my code as consistently running 1-2 ms slower than yours, even for single-word search inputs where we couldn't attribute this difference to the algorithm change. So I tried replacing the all/any functions with for-else loops, which to my surprise measured 1-2 ms faster than the existing code, even for multi-word cases. So I pushed that optimization.
Findings attached and as noted, this measures only the search so negates the network call latency, which far outweighs the search anyway.
138