discovery client

Node.js client for discovering Opentrons robots on the network

overview

@opentrons/discovery-client provides the ability to discover Opentrons robots on your network using MDNS.

api

const DiscoveryClient = require('@opentrons/discovery-client')

DiscoveryClientFactory(options?: Options): DiscoveryClient

Creates a new DiscoveryClient.

const options = {
  nameFilter: ['opentrons'],
  portFilter: [31950],
  pollInterval: 5000,
  candidates: [{ip: '[fd00:0:cafe:fefe::1]', port: 31950}, 'localhost'],
}

const client = DiscoveryClientFactory(options)

The discovery client is an Event Emitter. In addition to the normal EventEmitter methods and properties, the client has:

type DiscoveryClient = {
  /** list of discovered robot services */
  services: Array<Service>,

  /** list of extra service addresses the client is polling */
  candidates: Array<Candidate>,

  /**
   * start searching for robots
   *
   * @return {DiscoveryClient} `this` for chaining
   */
  start: () => DiscoveryClient,

  /**
   * stop searching for robots
   *
   * @return {DiscoveryClient} `this` for chaining
   */
  stop: () => DiscoveryClient,

  /**
   * adds an IP address to tracking manually (outside of MDNS discovery)
   *
   * @param {string} ip IP address to add to the tracking list
   * @param {number} port Optional port
   * @return {DiscoveryClient} `this` for chaining
   */
  add: (ip: string, port?: number) => DiscoveryClient,

  /**
   * removes a robot from the discovery list / tracking
   * (note: if a robot is and healthy, it will probably end up back on the list
   * automatically; this method is mostly for removing unhealthy robots)
   *
   * @param {string} name the name of the robot to remove
   * @return {DiscoveryClient} `this` for chaining
   */
  remove: (name: string) => DiscoveryClient

  /**
   * sets the health polling interval
   *
   * @param {?number} interval health polling interval; 0 sets to default
   * @return {DiscoveryClient} `this` for chaining
   */
  setPollInterval: (interval: number) => DiscoveryClient
}

type Service = {
  /** unique robot name */
  name: string,

  /** possible ip address (null if an IP conflict occurred) */
  ip: ?string,

  /** service port (deafult 31950) */
  port: number,

  /** IP address (if known) is a link-local address */
  local: ?boolean,

  /** health status of the API server (null if not yet determined) */
  ok: ?boolean,

  /** health status of the update server (null if not yet determined) */
  serverOk: ?boolean,

  /** whether robot is advertising over MDNS (null if not yet determined) */
  advertising: ?boolean,

  /** last good health response */
  health: ?{
    name: string,
    api_version: string,
    fw_version: string,
    system_version?: string,
    logs?: Array<string>,
  },

  /** last good update server health response */
  serverHealth: ?{
    name: string,
    apiServerVersion: string,
    updateServerVersion: string,
    smoothieVersion: string,
    systemVersion: string,
  },
}

type Candidate = {
  /** ip address */
  ip: string,

  /** service port (31950 if unspecified) */
  port: ?number,
}

discovery.client takes an optional options parameter:

type Options = {
  /**
   * interval (ms) at which to poll an IP to find a robot
   * default: 5000
   */
  pollInterval?: number,

  /**
   * list of discovered robots to pre-populate the discovered list
   * default: {}
   */
  services?: Array<Service>,

  /**
   * list of extra IP addresses to add to the search list
   * default: []
   */
  candidates?: Array<string | Candidate>,

  /**
   * RegExps or strings to filter services by name
   * default: ''
   */
  nameFilter?: Array<string | RegExp>,

  /**
   * RegExps or strings to filter services by IP address
   * default: ''
   */
  ipFilter?: Array<string | RegExp>,

  /**
   * array of numbers to filter services by port
   * the default port of 31950 is always included
   * default: []
   */
  portFilter?: Array<number>,

  /** optional logger */
  logger?: {
    ['error' | 'warn' | 'info' | 'debug']: (message: string, meta: {}) => void,
  },
}

If you need access to the DiscoveryClient class itself for some reason:

import {DiscoveryClient} from '@opentrons/discovery-client'

const client = new DiscoveryClient({})

events

import {SERVICE_EVENT, SERVICE_REMOVED_EVENT} from '@opentrons/discovery-client'

client.on(SERVICE_EVENT, (data) => console.log('service added/updated', data))
client.on(SERVICE_REMOVED_EVENT, (data) => console.log('service removed', data))
client.on('error', (error) => console.error(error)

| event name | data | description | | ---------------- | ---------------- | ---------------------------------------------- | | service | Array<Service> | Services added or updated in the services list | | serviceRemoved | Array<Service> | Services removed from the services list | | error | Error | MDNS encountered an error |

cli

@opentrons/discovery includes a simple CLI for convenience. After building the project by running make -C discovery-client, it is accessible through the node_modules directory of the monorepo. Launch it with:

# prereq: build project before first run or after changes
make -C discovery-client

# run via yarn run
yarn run discovery [command] [options]

# run via npx
npx discovery [command] [options]

# run from node_modules directly
node_modules/.bin/discovery [command] [options]

global options

The CLI's global options are almost completely the same as the API's options, with the addition of logLevel:

| flag | description | default | example | | -------------------- | ------------------------- | -------- | ---------------- | | -p, --pollInterval | see pollInterval option | 1000 | -p 500 | | -c, --candidates | see candidates option | [] | -c localhost | | -n, --nameFilter | see nameFilter option | [] | -n opentrons | | -i, --ipFilter | see ipFilter option | [] | -i 169.254 | | -a, --portFilter | see portFilter option | [] | -a 31951 31952 | | -l, --logLevel | log level for printout | 'info' | -l debug |

discovery (browse) [options]

Print out robots as it discovers them.

# example: browse for robots, including at localhost
discovery browse -c localhost

# browse is the default command, so you can leave the "browse" out
discovery --nameFilter moon

discovery find [name] [options]

Find the first robot you can, optionally specifying name or any other global options, and print out the IP address to stdout (bypassing any log level settings).

# example: find a specific robot
discovery find opentrons-moon-moon

# example: find the IP address of a link-local wired robot
discovery find --ipFilter 169.254

# example: find the IP address of a wired robot that may be IPv4 or IPv6
# (IPv6 means legacy non-mDNS wired configuration)
discovery find -i "169.254" "fd00" -c "[fd00:0:cafe:fefe::1]"

command specific options

| flag | description | default | example | | --------------- | ---------------------------- | ------- | ---------- | | -t, --timeout | How long to wait for a robot | 5000 | -t 10000 |

discovery-ssh [name] [options]

Calls discovery find and using the output to SSH into the robot it finds. Takes all the same arguments and options as discovery find.

discovery-ssh is a Bash script, so it must be called from a command line with Bash available.

# example: SSH into a link-local wired robot
discovery-ssh --ipFilter 169.254

# example: SSH into any wired robot, including legacy wired configuration
discovery-ssh -i "169.254" "fd00" -c "[fd00:0:cafe:fefe::1]"