simon/ocaml-containers
1
0
Fork 0
mirror of https://github.com/c-cube/ocaml-containers.git synced 2025-12-08 04:05:30 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

functor interface to Levenshtein automaton/index

Browse source
This commit is contained in:
Simon Cruanes 2014-03-05 16:49:10 +01:00
parent c5473857f8
commit b6310ae17d
2 changed files with 576 additions and 549 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1007
levenshtein.ml
View file

File diff suppressed because it is too large Load diff

118
levenshtein.mli
View file

@ -31,34 +31,19 @@ We take inspiration from
http://blog.notdot.net/2010/07/Damn-Cool-Algorithms-Levenshtein-Automata http://blog.notdot.net/2010/07/Damn-Cool-Algorithms-Levenshtein-Automata
for the main algorithm and ideas. However some parts are adapted *) for the main algorithm and ideas. However some parts are adapted *)
(** {2 Automaton} *) (** {2 Abstraction over Strings} *)
type 'a automaton module type STRING = sig
(** Levenshtein automaton for characters of type 'a *) type char_
type t
val of_array : ?compare:('a -> 'a -> int) -> limit:int -> 'a array -> 'a automaton val of_list : char_ list -> t
(** Build an automaton from an array, with a maximal distance [limit] *) val get : t -> int -> char_
val length : t -> int
val compare_char : char_ -> char_ -> int
end
val of_list : ?compare:('a -> 'a -> int) -> limit:int -> 'a list -> 'a automaton (** {2 Continuation list} *)
(** Build an automaton from a list, with a maximal distance [limit] *)
val of_string : limit:int -> string -> char automaton
(** Automaton for the special case of strings *)
val debug_print : out_channel -> char automaton -> unit
(** Output the automaton on the given channel. Only for string automata. *)
val match_with : 'a automaton -> 'a array -> bool
(** [match_with a s] matches the string [s] against [a], and returns
[true] if the distance from [s] to the word represented by [a] is smaller
than the limit used to build [a] *)
val match_with_string : char automaton -> string -> bool
(** Specialized version of {!match_with} for strings *)
(** {6 Index for one-to-many matching} *)
(** Continuation list *)
type 'a klist = type 'a klist =
[ [
| `Nil | `Nil
@ -68,49 +53,66 @@ type 'a klist =
val klist_to_list : 'a klist -> 'a list val klist_to_list : 'a klist -> 'a list
(** Helper. *) (** Helper. *)
module Index(X : Map.OrderedType) : sig (** {2 Signature} *)
type key = X.t
type 'b t module type S = sig
(** Index that maps [key] strings to values of type 'b. Internally it is type char_
based on a trie. *) type string_
val empty : 'b t (** {6 Automaton} *)
(** Empty index *)
val is_empty : _ t -> bool type automaton
(** Levenshtein automaton *)
val add : 'b t -> key array -> 'b -> 'b t val of_string : limit:int -> string_ -> automaton
(** Add a char array to the index. If a value was already present (** Build an automaton from an array, with a maximal distance [limit] *)
for this array it is replaced. *)
val remove : 'b t -> key array -> 'b -> 'b t val of_list : limit:int -> char_ list -> automaton
(** Remove a char array from the index. *) (** Build an automaton from a list, with a maximal distance [limit] *)
val retrieve : limit:int -> 'b t -> key array -> 'b klist val debug_print : (out_channel -> char_ -> unit) ->
(** Lazy list of objects associated to strings close to out_channel -> automaton -> unit
the query string *) (** Output the automaton on the given channel. *)
val of_list : (key array * 'b) list -> 'b t val match_with : automaton -> string_ -> bool
(** [match_with a s] matches the string [s] against [a], and returns
[true] if the distance from [s] to the word represented by [a] is smaller
than the limit used to build [a] *)
val to_list : 'b t -> (key array * 'b) list (** {6 Index for one-to-many matching} *)
(* TODO sequence/iteration functions *) module Index : sig
type 'b t
(** Index that maps strings to values of type 'b. Internally it is
based on a trie. *)
val empty : 'b t
(** Empty index *)
val is_empty : _ t -> bool
val add : 'b t -> string_ -> 'b -> 'b t
(** Add a char array to the index. If a value was already present
for this array it is replaced. *)
val remove : 'b t -> string_ -> 'b -> 'b t
(** Remove a string from the index. *)
val retrieve : limit:int -> 'b t -> string_ -> 'b klist
(** Lazy list of objects associated to strings close to the query string *)
val of_list : (string_ * 'b) list -> 'b t
val to_list : 'b t -> (string_ * 'b) list
(* TODO sequence/iteration functions *)
end
end end
(** Specific case for strings *) module Make(Str : STRING) : S
module StrIndex : sig with type string_ = Str.t
include module type of Index(Char) and type char_ = Str.char_
val add_string : 'b t -> string -> 'b -> 'b t include S with type char_ = char and type string_ = string
(** Add a string to a char index *)
val remove_string : 'b t -> string -> 'b -> 'b t val debug_print : out_channel -> automaton -> unit
(** Remove a string from a char index *)
val retrieve_string : limit:int -> 'b t -> string -> 'b klist
val of_str_list : (string * 'b) list -> 'b t
val to_str_list : 'b t -> (string * 'b) list
end