This is definitely not meant to be used by anyone. Just some example.
Given a wordlist seperated by \n, this will attempt to traverse through the list and ZADD [1] each item to the index with an initial frequency of 1. If the command EXISTS [2] returns true for the key, this step is redundant and will be skipped.
ZADD index 0 foo:1
The autocomplete function works by executing a range query using the command ZRANGEBYLEX [3] with the string provided, which returns all elements in the set sorted lexicographically between the given range. As an example, if the query provided is foo, we want to return all the results starting with foo. In this case the range would have a min of [foo with a max of [foo\xff (trailing byte).
ZRANGEBYLEX index "[foo" "[foo\xff"
This will then check if the initial search prefix was a 1:1 match with the result, if so we ZREM [4] the value and ZADD it back with an incremented frequency of 1, making the new value foo:2.
- [1] https://redis.io/commands/zadd
- [2] https://redis.io/commands/exists
- [3] https://redis.io/commands/zrangebylex
- [4] https://redis.io/commands/zrem
$ go run main.go \
--query="foo" \
--limit=10 \
--list="path_to_wordlist.txt" \
--port=6379
$ go test ./... --bench=.
BenchmarkReadWordListFunction-8 100 13850440 ns/op
BenchmarkHandleResultValuesFunction-8 20000000 74.8 ns/op
BenchmarkLexicographicalOrderFunction-8 20000 68123 ns/op
BenchmarkAddToListFunction-8 10000 118421 ns/op
BenchmarkCreateQueryFunction-8 5000000 262 ns/op
BenchmarkKeyExistsFunction-8 30000 55203 ns/op
BenchmarkHandleExactMatchFrequencyFunction-8 20000 89283 ns/op