package irmin

Irmin public API.

Irmin is a library to design and use persistent stores with built-in snapshot, branching and reverting mechanisms. Irmin uses concepts similar to Git but it exposes them as a high level library instead of a complex command-line frontend. It features a bidirectional Git backend, where an application can read and persist its state using the Git format, fully-compatible with the usual Git tools and workflows.

Irmin is designed to use a large variety of backends. It is written in pure OCaml and does not depend on external C stubs; it is thus very portable and aims to run everywhere, from Linux to browser and MirageOS unikernels.

Consult the basics and Examples of use for a quick start. See also the documentation for the unix backends.

Release 2.0.0 - %%HOMEPAGE%%

Preliminaries

Low-level Stores

An Irmin store is automatically built from a number of lower-level stores, each implementing fewer operations, such as content-addressable and atomic-write stores. These low-level stores are provided by various backends.

User-Defined Contents

High-level Stores

An Irmin store is a branch-consistent store where keys are lists of steps.

An example is a Git repository where keys are filenames, i.e. lists of '/'-separated strings. More complex examples are structured values, where steps might contain first-class field accessors and array offsets.

Irmin provides the following features:

• Support for fast clones, branches and merges, in a fashion very similar to Git.
• Efficient staging areas for fast, transient, in-memory operations.
• Fast synchronization primitives between remote stores, using native backend protocols (as the Git protocol) when available.

Synchronization

Examples

These examples are in the examples directory of the distribution.

Syncing with a remote

A simple synchronization example, using the Git backend and the Sync helpers. The code clones a fresh repository if the repository does not exist locally, otherwise it performs a fetch: in this case, only the missing contents are downloaded.

open Lwt.Infix

module S = Irmin_unix.Git.FS.KV(Irmin.Contents.String)
module Sync = Irmin.Sync(S)
let config = Irmin_git.config "/tmp/test"

let upstream =
  if Array.length Sys.argv = 2 then (Uri.of_string (Store.remote Sys.argv.(1)))
  else (Printf.eprintf "Usage: sync [uri]\n%!"; exit 1)

let test () =
  S.Repo.v config >>= S.master
  >>= fun t  -> Sync.pull_exn t upstream `Set
  >>= fun () -> S.get t ["README.md"]
  >|= fun r  -> Printf.printf "%s\n%!" r

let () = Lwt_main.run (test ())

Mergeable logs

We will demonstrate the use of custom merge operators by defining mergeable debug log files. We first define a log entry as a pair of a timestamp and a message, using the combinator exposed by Irmin.Type:

module Entry : sig
  include Irmin.Type.S
  val v: string -> t
  val timestamp: t -> int
end = struct

  type t = { timestamp: int; message : string; }

  let compare x y = compare x.timestamp y.timestamp

  let time = ref 0

  let v message =
    incr time;
    { timestamp = !time; message }

  let timestamp t = t.timestamp

  let pp ppf { timestamp; message } =
    Fmt.pf ppf "%04d: %s" timestamp message

  let of_string str =
    match String.split_on_char '\t' str with
    | [] -> Error (`Msg ("invalid entry: " ^ str))
    | ts :: msg_sects ->
      let message = String.concat "\t" msg_sects in
      try Ok { timestamp = int_of_string ts; message }
      with Failure e -> Error (`Msg e)

  let t =
    let open Irmin.Type in
    record "entry" (fun t32 message -> { timestamp = Int32.to_int t32; message })
    |+ field "timestamp" int32  (fun t -> Int32.of_int t.timestamp)
    |+ field "message"   string (fun t -> t.message)
    |> sealr

  let t = Irmin.Type.like ~cli:(pp, of_string) ~compare t

end

A log file is a list of entries (one per line), ordered by decreasing order of timestamps. The 3-way merge operator for log files concatenates and sorts the new entries and prepend them to the common ancestor's ones.

(* A log file *)
module Log: sig
  include Irmin.Contents.S
  val add: t -> Entry.t -> t
  val empty: t
end = struct

  type t = Entry.t list

  let empty = []

  let pp ppf l = List.iter (Fmt.pf ppf "%a\n" Entry.pp ) (List.rev l)

  let of_string str =
    let lines = String.cuts ~empty:false ~sep:"\n" str in
    try
      List.fold_left (fun acc l ->
          match Entry.of_string l with
          | Ok x           -> x :: acc
          | Error (`Msg e) -> failwith e
        ) [] lines
      |> fun l -> Ok l
    with Failure e ->
      Error (`Msg e)

  let t = Irmin.Type.(list Entry.t)
  let t = Irmin.Type.like' ~cli:(pp, of_string) t

  let timestamp = function
    | [] -> 0
    | e :: _ -> Entry.timestamp e

  let newer_than timestamp file =
    let rec aux acc = function
      | [] -> List.rev acc
      | h:: _ when Entry.timestamp h <= timestamp -> List.rev acc
      | h::t -> aux (h::acc) t
    in
    aux [] file

  let merge ~old t1 t2 =
    let open Irmin.Merge.Infix in
    old () >>=* fun old ->
    let old = match old with None -> [] | Some o -> o in
    let ts = timestamp old in
    let t1 = newer_than ts t1 in
    let t2 = newer_than ts t2 in
    let t3 = List.sort Entry.compare (List.rev_append t1 t2) in
    Irmin.Merge.ok (List.rev_append t3 old)

  let merge = Irmin.Merge.(option (v t merge))

  let add t e = e :: t

end 

Note: The serialisation primitives used in that example are not very efficient in this case as they parse the file every time. For real usage, you would write buffered versions of Log.pp and Log.of_string.

To persist the log file on disk, we need to choose a backend. We show here how to use the on-disk Git backend on Unix.

(* Build an Irmin store containing log files. *)
module S = Irmin_unix.Git.FS.KV(Log)

(* Set-up the local configuration of the Git repository. *)
let config = Irmin_git.config ~bare:true "/tmp/irmin/test"

(* Set-up the commit info function *)
let info fmt = Irmin_unix.info ~author:"logger" fmt

We can now define a toy example to use our mergeable log files.

open Lwt.Infix

(* Name of the log file. *)
let file = [ "local"; "debug" ]

(* Read the entire log file. *)
let read_file t =
  S.find t file >|= function
  | None   -> []
  | Some l -> l

(* Persist a new entry in the log. *)
let log t fmt =
  Fmt.kstrf (fun message ->
      read_file t >>= fun logs ->
      let logs = Log.add logs (Entry.v message) in
      S.set t (info "Adding a new entry") file logs
    ) fmt

let () =
  Lwt_main.run begin
    S.Repo.v config >>= S.master
    >>= fun t  -> log t "Adding a new log entry"
    >>= fun () -> Irmin.clone_force ~src:t ~dst:"x"
    >>= fun x  -> log x "Adding new stuff to x"
    >>= fun () -> log x "Adding more stuff to x"
    >>= fun () -> log x "More. Stuff. To x."
    >>= fun () -> log t "I can add stuff on t also"
    >>= fun () -> log t "Yes. On t!"
    >>= fun () -> S.merge (info "Merging x into t") x ~into:t
    >|= function Ok () -> () | Error _ -> failwith "merge conflict!"
  end

Helpers

Backends

API to create new Irmin backends. A backend is an implementation exposing either a concrete implementation of S or a functor providing S once applied.

There are two ways to create a concrete Irmin.S implementation:

• Make creates a store where all the objects are stored in the same store, using the same internal keys format and a custom binary format based on bin_prot, with no native synchronization primitives: it is usually what is needed to quickly create a new backend.
• Make_ext creates a store with a deep embedding of each of the internal stores into separate store, with total control over the binary format and using the native synchronization protocols when available.