4 ms·
Good ending with wtf. But it says 24% are better... It would be interesting to know how the ranking works.
by bambax 4y ago
Good ending with wtf.
But it says 24% are better... It would be interesting to know how the ranking works.
- drcode 4y agoHere's the core code, if you really want to know: (defn score [s] (let [s1 (st/lower-case s) s2 (remove-nonstandard s1) s3 (remove-numbers s2) splits (split-words s3) all-words? (= (count s3) (count (apply str splits)))] (cond-> {} all-words? (assoc :all-words 0) true (assoc :num (* (count s) (if all-words? -60 -100))) (names s3) (assoc :names 1000) (and (dictionary s3) (not (names s3))) (assoc :dictionary 1000) (not= (count s2) (count s3)) (assoc :numbers -500) (not= (count s1) (count s2)) (assoc :nonstandard -500))))
- drcode 4y agoI think the percentile code doesn't attempt to accurately handle identical scores and will give sloppy results in that case, so even if 24% have the same score it will report 24%