Skip to content

Instantly share code, notes, and snippets.

@gsg
Created November 11, 2014 13:07
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save gsg/936d74c3fbc1f75ad623 to your computer and use it in GitHub Desktop.
Save gsg/936d74c3fbc1f75ad623 to your computer and use it in GitHub Desktop.
(*
* GADT friendly hash tables.
*)
module type GadtHashEq = sig
type 'a key
type 'a value
val eq_assoc : 'a key -> 'b key -> 'b value -> 'a value option
val hash : 'a key -> int
end
module Make (O : GadtHashEq) : sig
type t
val create : unit -> t
val clear : t -> unit
val length : t -> int
val add : t -> 'a O.key -> 'a O.value -> unit
val alter : t -> 'a O.key -> ('a O.value -> 'a O.value) -> 'a O.value -> unit
val mem : t -> 'a O.key -> bool
val find : t -> 'a O.key -> 'a O.value option
val find_exn : t -> 'a O.key -> 'a O.value
val find_default : t -> 'a O.key -> 'a O.value -> 'a O.value
type iter_arg = { f : 'a . 'a O.key -> 'a O.value -> unit }
val iter : iter_arg -> t -> unit
type 'a fold_arg = { f : 'b . 'b O.key -> 'b O.value -> 'a -> 'a }
val fold : 'a fold_arg -> t -> 'a -> 'a
type info = {
max_chain_length : int;
chain_lengths : int array;
}
val info : t -> info
end = struct
type chain =
| Nil : chain
| Cons : 'a O.key * 'a O.value * chain -> chain
type t = {
mutable len : int;
mutable mask : int;
mutable data : chain array;
}
let create () = { len = 0; mask = 0; data = [|Nil|] }
let clear tbl =
tbl.len <- 0;
tbl.mask <- 0;
tbl.data <- [|Nil|]
let length tbl = tbl.len
let resize tbl newlen =
let newdata = Array.make newlen Nil in
let newmask = newlen - 1 in
let rec rehash_chain = function
| Nil -> ()
| Cons (k, v, rest) ->
let i = O.hash k land newmask in
newdata.(i) <- Cons (k, v, newdata.(i));
rehash_chain rest in
Array.iter rehash_chain tbl.data;
tbl.data <- newdata;
tbl.mask <- newmask
let is_pow2 n = n land (n - 1) = 0
let add tbl k v =
let len = succ tbl.len in
if is_pow2 len then resize tbl (Array.length tbl.data * 2);
let i = O.hash k land tbl.mask in
tbl.data.(i) <- Cons (k, v, tbl.data.(i));
tbl.len <- len
let find : type a . t -> a O.key -> a O.value option =
fun tbl key ->
let rec loop = function
| Nil -> None
| Cons (k, v, rest) ->
match O.eq_assoc key k v with
| None -> loop rest
| x -> x in
loop tbl.data.(O.hash key land tbl.mask)
let find_exn tbl key =
match find tbl key with
| Some x -> x
| None -> raise Not_found
let find_default tbl key default =
match find tbl key with
| Some x -> x
| None -> default
let mem tbl key =
match find tbl key with
| Some _ -> true
| None -> false
let alter : type a . t -> a O.key -> (a O.value -> a O.value) -> a O.value -> unit =
fun tbl key f init ->
let rec loop = function
| Nil -> Cons (key, init, Nil)
| Cons (k, v, rest) ->
match O.eq_assoc key k v with
| None -> Cons (k, v, loop rest)
| Some v' -> Cons (key, f v', rest) in
let i = O.hash key land tbl.mask in
tbl.data.(i) <- loop tbl.data.(i)
type iter_arg = { f : 'a . 'a O.key -> 'a O.value -> unit }
let iter : iter_arg -> t -> unit =
fun f tbl ->
let rec iter_chain = function
| Nil -> ()
| Cons (k, v, rest) -> f.f k v; iter_chain rest in
Array.iter iter_chain tbl.data
type 'a fold_arg = { f : 'b . 'b O.key -> 'b O.value -> 'a -> 'a }
let fold : 'a fold_arg -> t -> 'a -> 'a =
fun f tbl init ->
let x = ref init in
let rec fold_chain = function
| Nil -> ()
| Cons (k, v, rest) -> x := f.f k v !x; fold_chain rest in
Array.iter fold_chain tbl.data;
!x
type info = {
max_chain_length : int;
chain_lengths : int array;
}
let info tbl =
let maxlen = ref 0 in
let rec chainlen len = function
| Nil -> len
| Cons (_, _, rest) -> chainlen (len + 1) rest in
let lengths = Array.map (chainlen 0) tbl.data in
Array.iter (fun l -> maxlen := max !maxlen l) lengths;
{
max_chain_length = !maxlen;
chain_lengths = lengths;
}
end
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment