scala-autocomplete-api

Run it locally

sbt run

Run it in docker

docker build -f Dockerfile . -t scala-autocomplete-api:latest

docker run -d -p 0.0.0.0:3030:3030 --name scala-autocomplete-api scala-autocomplete-api

Curl to test

curl -H "Content-Type: application/json" -X POST -d '{"term":"foo"}' http://0.0.0.0:8080/autocomplete/wfst

What is this about?

I wanted to compare endpoints response time of Lucene's WFST - based autocomplete and Uncompressed Prefix Tree - hashMap based autocomplete

Comparison results:

I run this api locally and tested with hey

• wfst response time: 0.0500s
• hashmap response time: 0.0457s
• for comparison hashmap-based autocomplete in rust: 0.0185s

More info

wfst-based endpoint

command:

hey -n 10000 -c 200 http://0.0.0.0:3030/autocomplete/wfst -H "Content-Type: application/json" -X POST -d '{"term":"ap"}'

output:

Summary:
  Total:	4.2287 secs
  Slowest:	3.9930 secs
  Fastest:	0.0014 secs
  Average:	0.0500 secs
  Requests/sec:	2364.8065

  Total data:	479808 bytes
  Size/request:	48 bytes

Response time histogram:
  0.001 [1]	|
  0.401 [9818]	|■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■
  0.800 [18]	|
  1.199 [21]	|
  1.598 [12]	|
  1.997 [20]	|
  2.396 [23]	|
  2.796 [17]	|
  3.195 [22]	|
  3.594 [20]	|
  3.993 [24]	|


Latency distribution:
  10% in 0.0041 secs
  25% in 0.0053 secs
  50% in 0.0074 secs
  75% in 0.0108 secs
  90% in 0.0176 secs
  95% in 0.0267 secs
  99% in 2.1708 secs

Details (average, fastest, slowest):
  DNS+dialup:	0.0002 secs, 0.0014 secs, 3.9930 secs
  DNS-lookup:	0.0000 secs, 0.0000 secs, 0.0000 secs
  req write:	0.0000 secs, 0.0000 secs, 0.0142 secs
  resp wait:	0.0496 secs, 0.0013 secs, 3.9769 secs
  resp read:	0.0000 secs, 0.0000 secs, 0.0010 secs

Status code distribution:
  [405]	9996 responses

Error distribution:
  [1]	Get "http://0.0.0.0:3030/autocomplete/wfst": read tcp 127.0.0.1:59308->127.0.0.1:3030: read: connection reset by peer
  [1]	Get "http://0.0.0.0:3030/autocomplete/wfst": read tcp 127.0.0.1:59310->127.0.0.1:3030: read: connection reset by peer
  [1]	Get "http://0.0.0.0:3030/autocomplete/wfst": read tcp 127.0.0.1:59311->127.0.0.1:3030: read: connection reset by peer
  [1]	Get "http://0.0.0.0:3030/autocomplete/wfst": read tcp 127.0.0.1:59312->127.0.0.1:3030: read: connection reset by peer

hashMap-based endpoint

command:

hey -n 10000 -c 200 http://0.0.0.0:3030/autocomplete/hashmap -H "Content-Type: application/json" -X POST -d '{"term":"ap"}'

output:

Summary:
  Total:	3.7984 secs
  Slowest:	3.6009 secs
  Fastest:	0.0016 secs
  Average:	0.0457 secs
  Requests/sec:	2632.6863

  Total data:	480000 bytes
  Size/request:	48 bytes

Response time histogram:
  0.002 [1]	|
  0.362 [9823]	|■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■
  0.721 [13]	|
  1.081 [21]	|
  1.441 [20]	|
  1.801 [20]	|
  2.161 [19]	|
  2.521 [21]	|
  2.881 [22]	|
  3.241 [18]	|
  3.601 [22]	|


Latency distribution:
  10% in 0.0046 secs
  25% in 0.0061 secs
  50% in 0.0084 secs
  75% in 0.0119 secs
  90% in 0.0164 secs
  95% in 0.0209 secs
  99% in 1.8262 secs

Details (average, fastest, slowest):
  DNS+dialup:	0.0003 secs, 0.0016 secs, 3.6009 secs
  DNS-lookup:	0.0000 secs, 0.0000 secs, 0.0000 secs
  req write:	0.0000 secs, 0.0000 secs, 0.0161 secs
  resp wait:	0.0452 secs, 0.0015 secs, 3.5780 secs
  resp read:	0.0000 secs, 0.0000 secs, 0.0016 secs

Status code distribution:
  [405]	10000 responses