@glion/mllp-transport

MLLP wire protocol primitives — framing, encoding, decoding, and base error types.

What it does

@glion/mllp-transport implements the Minimal Lower Layer Protocol (MLLP) framing rules used to carry HL7v2 messages over a byte stream. It exports an encode() function that wraps a message in MLLP framing (<VT> message <FS><CR>), a decode() function for single-frame inputs, and a createDecoderStream() TransformStream that turns a chunked TCP byte stream into one decoded HL7v2 message per emitted value. It also exports the base MllpError class shared across the MLLP package family and the framing-specific TransportError subclass. This package has no transport implementation of its own — bring your own TCP, TLS, or WebSocket. @glion/mllp (server) and @glion/mllp-client build on top of it.

Install

npm install @glion/mllp-transport

Use

Encode a single HL7v2 message and write it to any byte sink:

import { encode } from "@glion/mllp-transport";

const message =
  "MSH|^~\\&|App|Fac|Recv|Fac|20240101120000||ADT^A01|MSG001|P|2.5.1\rPID|1||12345";

const frame = encode(message); // Uint8Array with VT prefix and FS CR suffix
socket.write(frame);

Decode a framed byte stream from a TCP socket:

import { Readable } from "node:stream";
import { createDecoderStream } from "@glion/mllp-transport";

const decoder = createDecoderStream({ maxMessageSize: 1_000_000 });
const reader = (Readable.toWeb(socket) as ReadableStream<Uint8Array>)
  .pipeThrough(decoder)
  .getReader();

while (true) {
  const { done, value: message } = await reader.read();
  if (done) {
    break;
  }
  console.log(message.text);
}

Catch transport-level framing errors:

import { ErrorCode, MllpError, TransportError } from "@glion/mllp-transport";

try {
  // ...
} catch (error) {
  if (error instanceof TransportError) {
    if (error.code === ErrorCode.MESSAGE_TOO_LARGE) {
      // Inbound frame exceeded maxMessageSize.
    }
  } else if (error instanceof MllpError) {
    // Some other MLLP-family error.
  }
}

API

encode(message, options?)

Encode a string or Uint8Array into an MLLP frame. Returns a Uint8Array with the start byte (0x0B) prefix and end sequence (0x1C 0x0D) suffix.

encodeMultiple(messages, options?)

Encode an array of messages into a single concatenated Uint8Array of frames.

decode(frame, options?)

Decode a single complete MLLP frame into a DecodedMessage. For chunked stream input, use createDecoderStream() instead.

createDecoderStream(options?)

Create a TransformStream<Uint8Array, DecodedMessage> that extracts complete MLLP frames from an arbitrarily-chunked byte stream.

The decoder offers two error modes, selected by what the onError callback does:

• Lenient (default) — the callback returns; the decoder logs the error and keeps scanning for the next valid frame. Right for receivers that process many frames over a long-lived connection.
• Fatal — the callback throws; the throw propagates out of transform() and errors the readable side of the TransformStream. Downstream for await ... of consumers reject with the thrown value. Right for one-shot exchanges (e.g. @glion/mllp-client) where any frame error invalidates the whole exchange.

// Lenient: log and keep going.
const lenient = createDecoderStream({
  maxMessageSize: 1_048_576,
  onError: (err) => log.warn(err),
});

// Fatal: error the stream on any frame error.
const fatal = createDecoderStream({
  maxMessageSize: 1_048_576,
  onError: (err) => {
    throw new Error(`framing failed: ${err.message}`, { cause: err });
  },
});

MLLPDecoderStream

Class-based variant of createDecoderStream, useful when a reusable instance is preferred over the factory.

Constants

Types

Errors

Part of Glion

@glion/mllp-transport is part of Glion, the application framework for HL7v2. See the Glion README for the full package catalog and architecture.