-
-
Save mseri/0cd0f8ed60a3ce0c24ba4c1e493eb29a to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
module Trie = struct | |
type t = { mutable count : int; children : t option Array.t } | |
let[@inline] empty () = { count = 0; children = Array.make 128 None } | |
let add buf pos len t = | |
let rec loop idx t = | |
if idx > pos + len - 1 then t.count <- t.count + 1 | |
else | |
let c = Char.lowercase_ascii (Bytes.get buf idx) in | |
let t' = | |
match t.children.(Char.code c) with | |
| None -> | |
let t'' = empty () in | |
t.children.(Char.code c) <- Some t''; | |
t'' | |
| Some c -> c | |
in | |
loop (idx + 1) t' | |
in | |
loop pos t | |
let to_assoc t = | |
let rec loop acc curr t = | |
let acc = | |
match (curr, t.count) with | |
| [], _ -> acc | |
| _, 0 -> acc | |
| curr, c -> (List.rev curr, c) :: acc | |
in | |
let r = ref acc in | |
Array.iteri | |
(fun idx t' -> | |
match t' with | |
| None -> () | |
| Some t' -> r := loop !r (Char.chr idx :: curr) t') | |
t.children; | |
!r | |
in | |
loop [] [] t | |
end | |
let[@inline] is_whitespace c = c = ' ' || c = '\n' | |
let rec skip_whitespace countwords pad ~lim i = | |
if i >= lim then (read_more_and_white [@tailcall]) countwords pad i | |
else if is_whitespace @@ Bytes.unsafe_get pad i then | |
(skip_whitespace [@tailcall]) countwords pad ~lim (i + 1) | |
else (read_next_word [@tailcall]) countwords pad ~lim i i | |
and read_next_word countwords pad ~lim i j = | |
if j >= lim then (read_more_and_next [@tailcall]) countwords pad i j | |
else if is_whitespace @@ Bytes.unsafe_get pad j then ( | |
Trie.add pad i (j - i) countwords; | |
(skip_whitespace [@tailcall]) countwords pad ~lim (j + 1)) | |
else (read_next_word [@tailcall]) countwords pad ~lim i (j + 1) | |
and read_more_and_next countwords pad i j = | |
let l = j - i in | |
Bytes.blit pad i pad 0 l; | |
let read = input stdin pad l (Bytes.length pad - l) in | |
let lim = l + read in | |
if lim = 0 then () else (read_next_word [@tailcall]) countwords pad ~lim 0 l | |
and read_more_and_white countwords pad i = | |
let lim = input stdin pad 0 (Bytes.length pad) in | |
if lim = 0 then () else (skip_whitespace [@tailcall]) countwords pad ~lim 0 | |
let[@inline] read_start countwords pad = | |
(read_more_and_next [@specialize]) countwords pad 0 0 | |
let () = | |
let t = Trie.empty () in | |
let pad = Bytes.make (64 * 1024) ' ' in | |
read_start t pad; | |
let assoc = Array.of_list (Trie.to_assoc t) in | |
Array.sort (fun (_, a) (_, b) -> Int.compare b a) assoc; | |
Array.iter | |
(fun (word, count) -> | |
List.iter print_char word; | |
print_char ' '; | |
print_int count; | |
print_newline ()) | |
assoc |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
(* Simpler IO doesn't make the timing worse *) | |
module Trie = struct | |
type t = { mutable count : int; children : t option Array.t } | |
let[@inline] empty () = { count = 0; children = Array.make 128 None } | |
let add s pos len t = | |
let rec loop idx t = | |
if idx > pos + len - 1 then t.count <- t.count + 1 | |
else | |
let c = Char.lowercase_ascii (String.get s idx) in | |
let t' = | |
match t.children.(Char.code c) with | |
| None -> | |
let t'' = empty () in | |
t.children.(Char.code c) <- Some t''; | |
t'' | |
| Some c -> c | |
in | |
loop (idx + 1) t' | |
in | |
loop pos t | |
let to_assoc t = | |
let rec loop acc curr t = | |
let acc = | |
match (curr, t.count) with | |
| [], _ -> acc | |
| _, 0 -> acc | |
| curr, c -> (List.rev curr, c) :: acc | |
in | |
let r = ref acc in | |
Array.iteri | |
(fun idx t' -> | |
match t' with | |
| None -> () | |
| Some t' -> r := loop !r (Char.chr idx :: curr) t') | |
t.children; | |
!r | |
in | |
loop [] [] t | |
end | |
let read_input t = | |
try | |
let rec loop () = | |
let line = input_line stdin in | |
let start = ref 0 in | |
let len = ref 0 in | |
for i = 0 to String.length line - 1 do | |
if line.[i] <> ' ' then incr len | |
else ( | |
Trie.add line !start !len t; | |
start := i + 1; | |
len := 0) | |
done; | |
Trie.add line !start !len t; | |
loop () | |
in | |
loop () | |
with End_of_file -> () | |
let () = | |
let t = Trie.empty () in | |
read_input t; | |
let assoc = Array.of_list (Trie.to_assoc t) in | |
Array.sort (fun (_, a) (_, b) -> Int.compare b a) assoc; | |
Array.iter | |
(fun (word, count) -> | |
List.iter print_char word; | |
print_char ' '; | |
print_int count; | |
print_newline ()) | |
assoc |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
module BytesHash = struct | |
include Bytes | |
let hash (k : t) = Hashtbl.hash k | |
end | |
module BytesHashtbl = Hashtbl.Make (BytesHash) | |
let lowercase_inplace pad i j = | |
for k = i to j - 1 do | |
Bytes.unsafe_set pad k @@ Char.lowercase_ascii @@ Bytes.unsafe_get pad k | |
done | |
let sub_and_add_word countwords source i len = | |
let word = Bytes.sub source i len in | |
match BytesHashtbl.find countwords word with | |
| v -> | |
BytesHashtbl.replace countwords word (v + 1) | |
| exception Not_found -> | |
BytesHashtbl.add countwords word 1 | |
let[@inline] is_whitespace c = c = ' ' || c = '\n' | |
let rec skip_whitespace countwords pad ~lim i = | |
if i >= lim then | |
(read_more_and_white[@taillcall]) countwords pad i | |
else if is_whitespace @@ Bytes.unsafe_get pad i then | |
(skip_whitespace[@tailcall]) countwords pad ~lim (i + 1) | |
else | |
(read_next_word[@tailcall]) countwords pad ~lim i i | |
and read_next_word countwords pad ~lim i j = | |
if j >= lim then | |
(read_more_and_next[@taillcall]) countwords pad i j | |
else if is_whitespace @@ Bytes.unsafe_get pad j then begin | |
sub_and_add_word countwords pad i (j-i); | |
(skip_whitespace[@taillcall]) countwords pad ~lim (j+1) | |
end else | |
(read_next_word[@taillcall]) countwords pad ~lim i (j + 1) | |
and read_more_and_next countwords pad i j = | |
let l = j - i in | |
Bytes.blit pad i pad 0 l; | |
let read = input stdin pad l (Bytes.length pad - l) in | |
let lim = l+read in | |
lowercase_inplace pad l lim; | |
if lim = 0 then () | |
else (read_next_word[@taillcall]) countwords pad ~lim 0 l | |
and read_more_and_white countwords pad i = | |
let lim = input stdin pad 0 (Bytes.length pad) in | |
lowercase_inplace pad 0 lim; | |
if lim = 0 then () | |
else (skip_whitespace[@taillcall]) countwords pad ~lim 0 | |
let[@inline] read_start countwords pad = | |
(read_more_and_next[@specialize]) countwords pad 0 0 | |
let[@inline] print_slot (w, c) = | |
print_bytes w; | |
print_char ' '; | |
print_int c; | |
print_newline () | |
let () = | |
let countwords = BytesHashtbl.create 33_000 in | |
let size = 64*1024 in | |
let pad = Bytes.make size ' ' in | |
read_start countwords pad ; | |
let a = Array.make (BytesHashtbl.length countwords) (Bytes.empty,0) in | |
let f k v i = a.(i) <- (k,v); i+1 in | |
let _ = BytesHashtbl.fold f countwords 0 in | |
Array.sort (fun (_,i) (_,j) -> Int.compare i j) a; | |
Array.iter print_slot a |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
module BytesHash = struct | |
include Bytes | |
let hash (k : bytes) = Hashtbl.hash k | |
end | |
module BytesHashtbl = Hashtbl.Make (BytesHash) | |
module Buf = struct | |
type t = { buf : Bytes.t; mutable pos : int; mutable len : int; size : int } | |
let create size = { buf = Bytes.create size; pos = 0; len = 0; size } | |
let[@inline] compress t = | |
if t.len <= 0 || t.pos = t.len then ( | |
t.pos <- 0; | |
t.len <- 0) | |
else if t.pos > 0 then ( | |
BytesLabels.unsafe_blit ~src:t.buf ~src_pos:t.pos ~dst:t.buf ~dst_pos:0 | |
~len:t.len; | |
t.pos <- 0) | |
let read_into t fd = | |
compress t; | |
let count = | |
UnixLabels.read fd ~buf:t.buf ~pos:(t.pos + t.len) ~len:(t.size - t.len) | |
in | |
t.len <- t.len + count; | |
for i = t.pos to t.len - 1 do | |
Bytes.unsafe_set t.buf i (Char.lowercase_ascii (Bytes.unsafe_get t.buf i)) | |
done; | |
count | |
let iter_words t tbl = | |
let prev = ref t.pos in | |
for i = t.pos to t.len - 1 do | |
(* We perform bounds check so its okay to call unsafe_get here *) | |
let c = Bytes.unsafe_get t.buf i in | |
if c = '\n' || c = ' ' then ( | |
let pos = !prev in | |
let len = i - !prev in | |
(if len > 0 then | |
let word = Bytes.sub t.buf pos len in | |
try incr (BytesHashtbl.find tbl word) | |
with Not_found -> BytesHashtbl.add tbl word (ref 1)); | |
prev := i + 1) | |
done; | |
t.len <- t.len - !prev; | |
t.pos <- !prev | |
end | |
let () = | |
let countwords = BytesHashtbl.create 33_000 in | |
let buf = Buf.create (64 * 1024) in | |
while Buf.read_into buf Unix.stdin > 0 do | |
Buf.iter_words buf countwords | |
done; | |
let arr = Array.make (BytesHashtbl.length countwords) (Bytes.empty, 0) in | |
let idx = ref 0 in | |
BytesHashtbl.iter | |
(fun word count -> | |
arr.(!idx) <- (word, !count); | |
incr idx) | |
countwords; | |
Array.sort (fun (_, x) (_, y) -> Int.compare y x) arr; | |
Array.iter | |
(fun (w, c) -> | |
print_bytes w; | |
print_char ' '; | |
print_int c; | |
print_newline ()) | |
arr |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment