A Ruby client for the Slack Web, RealTime Messaging and Events APIs. Comes with a handy command-line client, too. If you are not familiar with these concepts, you might want to watch this video.
- Useful to Me?
- Stable Release
- Installation
- Usage
- Create a New Bot Integration
- Use the API Token
- Global Settings
- Web Client
- Get Channel Info
- Get User Info
- Search for a User
- RealTime Client
- Configuring Slack::RealTime::Client
- Connection Methods
- Combining RealTime and Web Clients
- Large Team Considerations
- Events API
- Message Parsing
- Command-Line Client
- History
- Contributing
- Copyright and License
- This library will let you send messages to Slack via the Web API, send and receive messages via the Real Time Messaging API and facilitate integration with the Events API.
- To respond to slash commands, interactive components or events, implement a web application using your favorite web framework and use this library to call the Slack Web API and to verify that events are coming from Slack.
- To build a bot using the Real Time Messaging API, use slack-ruby-bot, which uses this library.
- To roll out a complete service using the Real Time Messaging API with Slack button integration to multiple teams, check out slack-ruby-bot-server, which is built on top of slack-ruby-bot, which uses this library.
You're reading the documentation for the next release of slack-ruby-client. Please see the documentation for the last stable release, v0.14.1 unless you're integrating with HEAD. See UPGRADING when upgrading from an older version.
Add to Gemfile.
gem 'slack-ruby-client'
If you're going to be using the RealTime client, add either async-websocket
, eventmachine
and faye-websocket
or celluloid-io
. See below for more information about concurrency. We recommend you use async-websocket
.
gem 'async-websocket'
Run bundle install
.
This is something done in Slack, under integrations. Create a new bot, and note its API token.
Slack.configure do |config|
config.token = ENV['SLACK_API_TOKEN']
end
This sets a global default token. You can also pass a token into the initializer of both Slack::Web::Client
and Slack::RealTime::Client
or configure those separately via Slack::Web::Config.configure
and Slack::RealTime::Config.configure
. The instance token will be used over the client type token over the global default.
The following global settings are supported via Slack.configure
.
setting | description |
---|---|
token | Slack API token. |
logger | An optional logger, defaults to ::Logger.new(STDOUT) at Logger::WARN level. |
The Slack Web API allows you to build applications that interact with Slack.
client = Slack::Web::Client.new
client.auth_test
Send messages with chat_PostMessage.
client.chat_postMessage(channel: '#general', text: 'Hello World', as_user: true)
See a fully working example in examples/hi_web.
List channels with channels_list.
channels = client.channels_list.channels
general_channel = channels.detect { |c| c.name == 'general' }
Upload a file with files_upload.
client.files_upload(
channels: '#general',
as_user: true,
file: Faraday::UploadIO.new('/path/to/avatar.jpg', 'image/jpeg'),
title: 'My Avatar',
filename: 'avatar.jpg',
initial_comment: 'Attached a selfie.'
)
You can use a channel ID or name (prefixed with #
) in all functions that take a :channel
argument. Lookup by name is not supported by the Slack API and the channels_id
method called invokes channels_list
in order to locate the channel ID.
client.channels_info(channel: 'C04KB5X4D') # calls channels_info
client.channels_info(channel: '#general') # calls channels_list followed by channels_info
You can use a user ID or name (prefixed with @
) in all functions that take a :user
argument. Lookup by name is not supported by the Slack API and the users_id
method called invokes users_list
in order to locate the user ID.
client.users_info(user: 'U092BDCLV') # calls users_info
client.users_info(user: '@dblock') # calls users_list followed by users_info
Constructs an in-memory index of users and searches it. If you want to use this functionality, add the picky gem to your project's Gemfile.
client.users_search(user: 'dblock')
Refer to the Slack Web API Method Reference for the list of all available functions.
You can configure the Web client either globally or via the initializer.
Slack::Web::Client.configure do |config|
config.user_agent = 'Slack Ruby Client/1.0'
end
client = Slack::Web::Client.new(user_agent: 'Slack Ruby Client/1.0')
The following settings are supported.
setting | description |
---|---|
token | Slack API token. |
user_agent | User-agent, defaults to Slack Ruby Client/version. |
proxy | Optional HTTP proxy. |
ca_path | Optional SSL certificates path. |
ca_file | Optional SSL certificates file. |
endpoint | Slack endpoint, default is https://slack.com/api. |
logger | Optional Logger instance that logs HTTP requests. |
timeout | Optional open/read timeout in seconds. |
open_timeout | Optional connection open timeout in seconds. |
default_page_size | Optional page size for paginated requests, default is 100. |
default_max_retries | Optional number of retries for paginated requests, default is 100. |
You can also pass request options, including timeout
and open_timeout
into individual calls.
client.channels_list(request: { timeout: 180 })
You can also control what proxy options are used by modifying the http_proxy
environment variable per Net::HTTP's documentation.
Note that Docker on OSX seems to incorrectly set the proxy, causing Faraday::ConnectionFailed, ERROR -- : Failed to open TCP connection to : (getaddrinfo: Name or service not known)
. You might need to manually unset http_proxy
in that case, eg. http_proxy="" bundle exec ruby ./my_bot.rb
.
The Web client natively supports cursor pagination for methods that allow it, such as users_list
. Supply a block and the client will make repeated requests adjusting the value of cursor
with every response. The default limit is set to 100 and can be adjusted via Slack::Web::Client.config.default_page_size
or by passing it directly into the API call.
all_members = []
client.users_list(presence: true, limit: 10) do |response|
all_members.concat(response.members)
end
all_members # many thousands of team members retrieved 10 at a time
When using cursor pagination the client will automatically pause and then retry the request if it runs into Slack rate limiting. (It will pause according to the Retry-After
header in the 429 response before retrying the request.) If it receives too many rate-limited responses in a row it will give up and raise an error. The default number of retries is 100 and can be adjusted via Slack::Web::Client.config.default_max_retries
or by passing it directly into the method as max_retries
.
You can also proactively avoid rate limiting by adding a pause between every paginated request with the sleep_interval
parameter, which is given in seconds.
all_members = []
client.users_list(presence: true, limit: 10, sleep_interval: 5, max_retries: 20) do |response|
# pauses for 5 seconds between each request
# gives up after 20 consecutive rate-limited responses
all_members.concat(response.members)
end
all_members # many thousands of team members retrieved 10 at a time
If a request fails, a Slack::Web::Api::Errors::SlackError
will be raised. The error message contains the error code. In case of multiple errors, the error codes are separated by commas. The original response is also accessible using the response
attribute.
If you exceed Slack’s rate limits, a Slack::Web::Api::Errors::TooManyRequestsError
will be raised instead.
In any other case, a Faraday::ClientError
will be raised. This may be the case if Slack is temporarily unavailable, for example.
The Real Time Messaging API is a WebSocket-based API that allows you to receive events from Slack in real time and send messages as user.
client = Slack::RealTime::Client.new
client.on :hello do
puts "Successfully connected, welcome '#{client.self.name}' to the '#{client.team.name}' team at https://#{client.team.domain}.slack.com."
end
client.on :message do |data|
case data.text
when 'bot hi' then
client.message(channel: data.channel, text: "Hi <@#{data.user}>!")
when /^bot/ then
client.message(channel: data.channel, text: "Sorry <@#{data.user}>, what?")
end
end
client.on :close do |_data|
puts "Client is about to disconnect"
end
client.on :closed do |_data|
puts "Client has disconnected successfully!"
end
client.start!
You can send typing indicators with typing
.
client.typing channel: data.channel
You can send a ping with ping
.
client.ping
By default, the RealTime client exposes and maintains a local store with the properties of rtm.start upon a successful connection.
property | description |
---|---|
url | A WebSocket Message Server URL. |
self | The authenticated bot user. |
team | Details on the authenticated user's team. |
users | A hash of user objects by user ID. |
channels | A hash of channel objects, one for every channel visible to the authenticated user. |
groups | A hash of group objects, one for every group the authenticated user is in. |
ims | A hash of IM objects, one for every direct message channel visible to the authenticated user. |
bots | Details of the integrations set up on this team. |
It also tracks changes, such as users being renamed, added or deleted, therefore client.users
is always up-to-date.
Tracking with a local store can be disabled with Slack::RealTime::Client.new(store_class: nil)
. Other stores are also available.
The default store that tracks all changes. By default the client will be connected using rtm_start
.
A smaller store that only stores and tracks information about the bot user, but not channels, users, groups, ims or bots. By default the client will be connected using rtm_connect
.
You can configure the RealTime client either globally or via the initializer.
Slack::RealTime::Client.configure do |config|
config.websocket_ping = 42
end
client = Slack::RealTime::Client.new(websocket_ping: 42)
The following settings are supported.
setting | description |
---|---|
token | Slack API token. |
websocket_ping | How long the socket can be idle before sending a ping message to confirm it's still connected, default is 30. |
websocket_proxy | Connect via proxy, include :origin and :headers . |
store_class | Local store class name, default is an in-memory Slack::RealTime::Stores::Store . |
start_method | Optional start method, either :rtm_start or :rtm_connect . |
start_options | Options to pass into rtm.start or rtm.connect , default is { request: { timeout: 180 } } . |
logger | Optional Logger instance that logs RealTime requests and socket data. |
Note that the RealTime client uses a Web client to obtain the WebSocket URL via rtm.start or rtm.connect. While token
and logger
options are passed down from the RealTime client, you may also configure Web client options via Slack::Web::Client.configure
as described above.
See a fully working example in examples/hi_real_time.
This setting determines how long the socket can be idle before sending a ping message to confirm it's still connected.
It's important to note that if a ping message was sent and no response was received within the amount of time specified in websocket_ping
; the client will attempt to reestablish it's connection to the message server.
To disable this feature; set websocket_ping
to 0.
The RealTime client uses either rtm.start or rtm.connect to open a connection. The former retrieves a lot of team information while the latter only serves connection purposes and is preferred. You should let the library choose the right method for you based on the store_class
used and override this behavior with start_method
when necessary.
Slack::RealTime::Client.configure do |config|
config.start_method = :rtm_start
end
Since the Web client is used to obtain the RealTime client's WebSocket URL, you can continue using the Web client in combination with the RealTime client.
client = Slack::RealTime::Client.new
client.on :message do |data|
case data.text
when 'bot hi' then
client.web_client.chat_postMessage(channel: data.channel, text: "Hi <@#{data.user}>!")
when /^bot/ then
client.web_client.chat_postMessage(channel: data.channel, text: "Sorry <@#{data.user}>, what?")
end
end
client.start!
See a fully working example in examples/hi_real_time_and_web.
The rtm.start
call downloads a large amount of data. For large teams, consider reducing the amount of unnecessary data downloaded with start_options
. You may also want to increase the default timeout of 180 seconds.
Slack::RealTime::Client.configure do |config|
# Return timestamp only for latest message object of each channel.
config.start_options[:simple_latest] = true
# Skip unread counts for each channel.
config.start_options[:no_unreads] = true
# Increase request timeout to 6 minutes.
config.start_options[:request][:timeout] = 360
end
See #134 for a discussion on this topic.
Slack::RealTime::Client
needs help from a concurrency library and supports Async, Faye::WebSocket with Eventmachine and Celluloid. It will auto-detect one or the other depending on the gems in your Gemfile, but you can also set concurrency explicitly.
Slack::RealTime.configure do |config|
config.concurrency = Slack::RealTime::Concurrency::Async
end
Use client.start_async
instead of client.start!
. A good example of such application is slack-bot-server.
client = Slack::RealTime::Client.new
client.start_async
This is the recommended library. Add async-websocket
to your Gemfile.
gem 'async-websocket'
See a fully working example in examples/hi_real_time_async_async.
Add the following to your Gemfile.
gem 'faye-websocket'
See a fully working example in examples/hi_real_time_async_eventmachine.
Add the following to your Gemfile.
gem 'celluloid-io', require: ['celluloid/current', 'celluloid/io']
See a fully working example in examples/hi_real_time_async_celluloid.
This library provides limited support for the Slack Events API.
You can configure Events support globally.
Slack::Events.configure do |config|
config.signing_secret = 'secret'
end
The following settings are supported.
setting | description |
---|---|
signing_secret | Slack signing secret, defaults is ENV['SLACK_SIGNING_SECRET'] . |
signature_expires_in | Signature expiration window in seconds, default is 300 . |
Slack signs its requests using a secret that's unique to your app. Verify incoming HTTP requests as follows.
slack_request = Slack::Events::Request.new(http_request)
slack_request.verify!
To specify secrets on a per-request basis:
Slack::Events::Request.new(http_request,
signing_secret: signing_secret,
signature_expires_in: signature_expires_in)
The verify!
call may raise Slack::Events::MissingSigningSecret
, Slack::Events::InvalidSignature
or Slack::Events::TimestampExpired
errors.
All text in Slack uses the same system of escaping: chat messages, direct messages, file comments, etc. Use Slack::Messages::Formatting to unescape incoming messages. This comes handy, for example, you want to treat all input to a real time bot as plain text.
Slack::Messages::Formatting.unescape('Hello & <world>'))
# => 'Hello & <world>'
Slack::Messages::Formatting.unescape('Hey <@U024BE7LH|bob>, did you see my file?'))
# => 'Hey @bob, did you see my file?'
Slack::Messages::Formatting.unescape('Hey <@U02BEFY4U>'))
# => 'Hey @U02BEFY4U'
Slack::Messages::Formatting.unescape('This message contains a URL <http://foo.com/>'))
# => 'This message contains a URL http://foo.com/'
Slack::Messages::Formatting.unescape('So does this one: <http://www.foo.com|www.foo.com>'))
# => 'So does this one: www.foo.com'
Slack::Messages::Formatting.unescape('<mailto:bob@example.com|Bob>'))
# => 'Bob'
Slack::Messages::Formatting.unescape('Hello <@U123|bob>, say hi to <!everyone> in <#C1234|general>'))
# => 'Hello @bob, say hi to @everyone in #general'
Slack::Messages::Formatting.unescape('Hello <@U123|bob> > file.txt'))
# => 'Hello @bob > file.txt'
Slack::Messages::Formatting.unescape('“hello”'))
# => '"hello"'
Slack::Messages::Formatting.unescape('‘hello’'))
# => "'hello'"
The slack command-line client returns JSON data from the Slack API.
$ slack --slack-api-token=[token] auth test
{"ok":true,"url":"...","team":"...","user":"...","team_id":"...","user_id":"..."}
export SLACK_API_TOKEN=...
$ slack chat postMessage --text="hello world" --channel="#general"
{"ok":true,"channel":"...","ts":"...","message":{"text":"hello world","username":"bot","type":"message","subtype":"bot_message","ts":"..."}}
$ slack channels id --channel=#general
{"ok":true,"channel":{"id":"C04KB5X4D"}}
$ slack channels info --channel=#general
{"ok":true,"channel":{"id":"C04KB5X4D","name":"general", ...}}
Combine with jq, a command-line JSON parser.
$ slack users list | jq '.members | map({(.id): .name})'
[
{
"U04KB5WQR": "dblock"
},
{
"U07518DTL": "rubybot"
}
]
See slack help
for a complete command-line reference.
This gem is based on slack-ruby-gem, but it more clearly separates the Web and RTM APIs, is more thoroughly tested and is in active development.
See CONTRIBUTING.
Copyright (c) 2015-2019, Daniel Doubrovkine, Artsy and Contributors.
This project is licensed under the MIT License.