Module Jsonm

module Jsonm: sig .. end
Non-blocking streaming JSON codec.

Jsonm is a non-blocking streaming codec to decode and encode the JSON data format. It can process JSON text without blocking on IO and without a complete in-memory representation of the data.

The uncut codec also processes whitespace and (non-standard) JSON with JavaScript comments.

Consult the data model, limitations and examples of use.

v1.0.0 - homepage

References




JSON data model


type lexeme = [ `Ae
| `As
| `Bool of bool
| `Float of float
| `Name of string
| `Null
| `Oe
| `Os
| `String of string ]
The type for JSON lexemes. `As and `Ae start and end arrays and `Os and `Oe start and end objects. `Name is for the member names of objects.

A well-formed sequence of lexemes belongs to the language of the json grammar:

  json = value 
object = `Os *member `Oe
member = (`Name s) value
 array = `As *value `Ae
 value = `Null / `Bool b / `Float f / `String s / object / array
A decoder returns only well-formed sequences of lexemes or `Errors are returned. The UTF-8, UTF-16, UTF-16LE and UTF-16BE encoding schemes are supported. The strings of decoded `Name and `String lexemes are however always UTF-8 encoded. In these strings, characters originally escaped in the input are in their unescaped representation.

An encoder accepts only well-formed sequences of lexemes or Invalid_argument is raised. Only the UTF-8 encoding scheme is supported. The strings of encoded `Name and `String lexemes are assumed to be immutable and must be UTF-8 encoded, this is not checked by the module. In these strings, the delimiter characters U+0022 and U+005C ('"', '\') aswell as the control characters U+0000-U+001F are automatically escaped by the encoders, as mandated by the standard.

val pp_lexeme : Format.formatter -> [< lexeme ] -> unit
pp_lexeme ppf l prints a unspecified non-JSON representation of l on ppf.

Decode


type error = [ `Expected of
[ `Aval of bool
| `Comment
| `Eoi
| `Json
| `Name
| `Name_sep
| `Omem of bool
| `Value ]
| `Illegal_BOM
| `Illegal_bytes of string
| `Illegal_escape of
[ `Lone_hi_surrogate of int
| `Lone_lo_surrogate of int
| `Not_esc_uchar of Uchar.t
| `Not_hex_uchar of Uchar.t
| `Not_lo_surrogate of int ]
| `Illegal_literal of string
| `Illegal_number of string
| `Illegal_string_uchar of Uchar.t
| `Unclosed of [ `As | `Comment | `Os | `String ] ]

The type for decoding errors.
val pp_error : Format.formatter -> [< error ] -> unit
pp_error e prints an unspecified UTF-8 representation of e on ppf.
type encoding = [ `UTF_16 | `UTF_16BE | `UTF_16LE | `UTF_8 ] 
The type for Unicode encoding schemes.
type src = [ `Channel of Pervasives.in_channel | `Manual | `String of string ] 
The type for input sources. With a `Manual source the client must provide input with Jsonm.Manual.src.
type decoder 
The type for JSON decoders.
val decoder : ?encoding:[< encoding ] -> [< src ] -> decoder
decoder encoding src is a JSON decoder that inputs from src. encoding specifies the character encoding of the data. If unspecified the encoding is guessed as suggested by the old RFC4627 standard.
val decode : decoder ->
[> `Await | `End | `Error of error | `Lexeme of lexeme ]
decode d is:

The Jsonm.Uncut.pp_decode function can be used to inspect decode results.

Note. Repeated invocation always eventually returns `End, even in case of errors.

val decoded_range : decoder -> (int * int) * (int * int)
decoded_range d is the range of characters spanning the last `Lexeme or `Error (or `White or `Comment for an Jsonm.Uncut.decode) decoded by d. A pair of line and column numbers respectively one and zero based.
val decoder_encoding : decoder -> encoding
decoder_encoding d is d's encoding.

Warning. If the decoder guesses the encoding, rely on this value only after the first `Lexeme was decoded.

val decoder_src : decoder -> src
decoder_src d is d's input source.

Encode


type dst = [ `Buffer of Buffer.t | `Channel of Pervasives.out_channel | `Manual ] 
The type for output destinations. With a `Manual destination the client must provide output storage with Jsonm.Manual.dst.
type encoder 
The type for JSON encoders.
val encoder : ?minify:bool -> [< dst ] -> encoder
encoder minify dst is an encoder that outputs to dst. If minify is true (default) the output is made as compact as possible, otherwise the output is indented. If you want better control on whitespace use minify = true and Jsonm.Uncut.encode.
val encode : encoder ->
[< `Await | `End | `Lexeme of lexeme ] -> [ `Ok | `Partial ]
encode e v is: For `Manual destinations, encoding `End always returns `Partial, the client should as usual use Jsonm.Manual.dst and continue with `Await until `Ok is returned at which point Jsonm.Manual.dst_rem e is guaranteed to be the size of the last provided buffer (i.e. nothing was written).

Raises. Invalid_argument if a non well-formed sequence of lexemes is encoded or if `Lexeme or `End is encoded after a `Partial encode.

val encoder_dst : encoder -> dst
encoder_dst e is e's output destination.
val encoder_minify : encoder -> bool
encoder_minify e is true if e's output is minified.

Manual sources and destinations


module Manual: sig .. end
Manual input sources and output destinations.

Uncut codec


module Uncut: sig .. end
Codec with comments and whitespace.

Limitations

Decode

Decoders parse valid JSON with the following limitations:

Position tracking assumes that each decoded Unicode scalar value has a column width of 1. The same assumption may not be made by the display program (e.g. for emacs' compilation mode you need to set compilation-error-screen-columns to nil).

The newlines LF (U+000A), CR (U+000D), and CRLF are all normalized to LF internally. This may have an impact in some corner `Error cases. For example the invalid escape sequence <U+005C,U+000D> in a string will be reported as being `Illegal_escape (`Not_esc_uchar 0x000A).

Encode

Encoders produce valid JSON provided the client ensures that the following holds.



Error recovery

After a decoding error, if best-effort decoding is performed. The following happens before continuing:



Examples

Trip

The result of trip src dst has the JSON from src written on dst.

let trip ?encoding ?minify
    (src : [`Channel of in_channel | `String of string])
    (dst : [`Channel of out_channel | `Buffer of Buffer.t])
  =
  let rec loop d e = match Jsonm.decode d with
  | `Lexeme _ as v -> ignore (Jsonm.encode e v); loop d e
  | `End -> ignore (Jsonm.encode e `End); `Ok
  | `Error err -> `Error (Jsonm.decoded_range d, err)
  | `Await -> assert false
  in
  let d = Jsonm.decoder ?encoding src in
  let e = Jsonm.encoder ?minify dst in
  loop d e
Using the `Manual interface, trip_fd does the same but between Unix file descriptors.
let trip_fd ?encoding ?minify
    (fdi : Unix.file_descr)
    (fdo : Unix.file_descr)
  =
  let rec encode fd s e v = match Jsonm.encode e v with `Ok -> ()
  | `Partial ->
      let rec unix_write fd s j l =
        let rec write fd s j l = try Unix.single_write fd s j l with
        | Unix.Unix_error (Unix.EINTR, _, _) -> write fd s j l
        in
        let wc = write fd s j l in
        if wc < l then unix_write fd s (j + wc) (l - wc) else ()
      in
      unix_write fd s 0 (Bytes.length s - Jsonm.Manual.dst_rem e);
      Jsonm.Manual.dst e s 0 (String.length s);
      encode fd s e `Await
  in
  let rec loop fdi fdo ds es d e = match Jsonm.decode d with
  | `Lexeme _ as v -> encode fdo es e v; loop fdi fdo ds es d e
  | `End -> encode fdo es e `End`Ok
  | `Error err -> `Error (Jsonm.decoded_range d, err)
  | `Await ->
      let rec unix_read fd s j l = try Unix.read fd s j l with
      | Unix.Unix_error (Unix.EINTR, _, _) -> unix_read fd s j l
      in
      let rc = unix_read fdi ds 0 (Bytes.length ds) in
      Jsonm.Manual.src d ds 0 rc; loop fdi fdo ds es d e
  in
  let ds = Bytes.create 65536 (* UNIX_BUFFER_SIZE in 4.0.0 *) in
  let es = Bytes.create 65536 (* UNIX_BUFFER_SIZE in 4.0.0 *) in
  let d = Jsonm.decoder ?encoding `Manual in
  let e = Jsonm.encoder ?minify `Manual in
  Jsonm.Manual.dst e es 0 (Bytes.length es);
  loop fdi fdo ds es d e

Member selection

The result of memsel names src is the list of string values of members of src that have their name in names. In this example, decoding errors are silently ignored.

let memsel ?encoding names
    (src : [`Channel of in_channel | `String of string])
  =
  let rec loop acc names d = match Jsonm.decode d with
  | `Lexeme (`Name n) when List.mem n names ->
      begin match Jsonm.decode d with
      | `Lexeme (`String s) -> loop (s :: acc) names d
      | _ -> loop acc names d
      end
  | `Lexeme _ | `Error _ -> loop acc names d
  | `End -> List.rev acc
  | `Await -> assert false
  in
  loop [] names (Jsonm.decoder ?encoding src)

Generic JSON representation

A generic OCaml representation of JSON text is the following one.

type json =
  [ `Null | `Bool of bool | `Float of float| `String of string
  | `A of json list | `O of (string * json) list ]
The result of json_of_src src is the JSON text from src in this representation. The function is tail recursive.
exception Escape of ((int * int) * (int * int)) * Jsonm.error

let json_of_src ?encoding
    (src : [`Channel of in_channel | `String of string])
  =
  let dec d = match Jsonm.decode d with
  | `Lexeme l -> l
  | `Error e -> raise (Escape (Jsonm.decoded_range d, e))
  | `End | `Await -> assert false
  in
  let rec value v k d = match v with
  | `Os -> obj [] k d  | `As -> arr [] k d
  | `Null | `Bool _ | `String _ | `Float _ as v -> k v d
  | _ -> assert false
  and arr vs k d = match dec d with
  | `Ae -> k (`A (List.rev vs)) d
  | v -> value v (fun v -> arr (v :: vs) k) d
  and obj ms k d = match dec d with
  | `Oe -> k (`O (List.rev ms)) d
  | `Name n -> value (dec d) (fun v -> obj ((n, v) :: ms) k) d
  | _ -> assert false
  in
  let d = Jsonm.decoder ?encoding src in
  try `JSON (value (dec d) (fun v _ -> v) d) with
  | Escape (r, e) -> `Error (r, e)
The result of json_to_dst dst json has the JSON text json written on dst. The function is tail recursive.
let json_to_dst ~minify
    (dst : [`Channel of out_channel | `Buffer of Buffer.t ])
    (json : json)
  =
  let enc e l = ignore (Jsonm.encode e (`Lexeme l)) in
  let rec value v k e = match v with
  | `A vs -> arr vs k e
  | `O ms -> obj ms k e
  | `Null | `Bool _ | `Float _ | `String _ as v -> enc e v; k e
  and arr vs k e = enc e `As; arr_vs vs k e
  and arr_vs vs k e = match vs with
  | v :: vs' -> value v (arr_vs vs' k) e
  | [] -> enc e `Ae; k e
  and obj ms k e = enc e `Os; obj_ms ms k e
  and obj_ms ms k e = match ms with
  | (n, v) :: ms -> enc e (`Name n); value v (obj_ms ms k) e
  | [] -> enc e `Oe; k e
  in
  let e = Jsonm.encoder ~minify dst in
  let finish e = ignore (Jsonm.encode e `Endin
  value json finish e