New Features
Sign in with Slack (OpenID Connect)
Sign in with Slack helps users log into your service using their Slack profile. The platform feature was recently upgraded to be compatible with the standard OpenID Connect specification. With this version of slack-sdk
, implementing the auth flow is much easier.
When you create a new Slack app, set the following user scopes:
oauth_config:
redirect_urls:
- https://{your-domain}/slack/oauth_redirect
scopes:
user:
- openid # required
- email # optional
- profile # optional
Check the Flask example in the repository. It does the following:
- Build the OpenID Connect comaptible authorize URL
slack_sdk.oauth.OpenIDConnectAuthorizeUrlGenerator
helps you easily do thisOAuthStateStore
is still available for generatingstate
parameter value (it's available fornonce
management too)
WebClient
can performopenid.connect.token
API calls with givencode
parameter
If you want to know the way with asyncio, check the Sanic app example in the same directory.
Built-in Retry Handlers
This version introduces a built-in retry functionalities to the following API clients:
slack_sdk.web.WebClient
slack_sdk.webhook.WebhookClient
slack_sdk.audit_logs.AuditLogsClient
slack_sdk.scim.SCIMClient
slack_sdk.web.async_client.AsyncWebClient
(aiohttp/asyncio compatible)slack_sdk.webhook.async_client.AsyncWebhookClient
(aiohttp/asyncio compatible)slack_sdk.audit_logs.async_client.AsyncAuditLogsClient
(aiohttp/asyncio compatible)slack_sdk.scim.async_client.AsyncSCIMClient
(aiohttp/asyncio compatible)
With the default settings, only slack_sdk.http_retry.builtin_handlers.ConnectionErrorRetryHandler
(AsyncConnectionErrorRetryHandler
for asyncio clients) with its default configuratio (=only one retry in the manner of exponential backoff and jitter) is enabled in the above clients. The retry handler retries if an API client encounters a connectivity-related failure (e.g., Connection reset by peer).
import os
from slack_sdk.web import WebClient
# Only ConnectionErrorRetryHandler is enabled
client = WebClient(token=os.environ["SLACK_BOT_TOKEN"])
Another built-in retry handler is the one that handles rate limited errors.
# --------------------------------
# Use the built-in handlers
# --------------------------------
client = WebClient(token=os.environ["SLACK_BOT_TOKEN"])
# This handler does retries when HTTP status 429 is returned
from slack_sdk.http_retry.builtin_handlers import RateLimitErrorRetryHandler
rate_limit_handler = RateLimitErrorRetryHandler(max_retry_count=1)
# Enable rate limited error retries as well
cient.retry_handlers.append(rate_limit_handler)
Creating your own ones is also quite simple. Defining a new class that inherits slack_sdk.http_retry.RetryHandler
and implements required methods (internals of can_retry / prepare_for_next_retry). Check the built-in ones' source code for learning how to properly implement.
# --------------------------------
# Create your own one
# --------------------------------
import socket
from typing import Optional
from slack_sdk.http_retry import (RetryHandler, RetryState, HttpRequest, HttpResponse)
from slack_sdk.http_retry.builtin_interval_calculators import (
BackoffRetryIntervalCalculator,
)
from slack_sdk.http_retry.jitter import RandomJitter
class MyRetryHandler(RetryHandler):
def _can_retry(
self,
*,
state: RetryState,
request: HttpRequest,
response: Optional[HttpResponse] = None,
error: Optional[Exception] = None
) -> bool:
# [Errno 104] Connection reset by peer
return (
error is not None and isinstance(error, socket.error) and error.errno == 104
)
# Customize the settings
my_handler = MyRetryHandler(
max_retry_count=2, # retry twice at maximum; the default is 1
interval_calculator=BackoffRetryIntervalCalculator( # exponential backoff and jitter is the default
backoff_factor=1.0, # 1, 2, 4, 8, 16 seconds later ...
jitter=RandomJitter(), # will add 0.0 -- 1.0 second to the backoff duration
),
)
client = WebClient(
token=os.environ["SLACK_BOT_TOKEN"],
retry_handlers=[rate_limit_handler, my_handler],
)
For asyncio apps, Async
prefixed corresponding modules are available. All the methods in those methods are async/await compatible. Check the source code and tests for more details.
Changes
- #1079 Fix #1078 Add Sign in with Slack (OpenID Connect) support - Thanks @seratch
- #1084 Fix #887 Enable automatic retry by a handy way - Thanks @seratch
- #1089 #1058 Add Slack Connect API support - Thanks @srajiang
- #1067 Fix #1065 by having lock for async Socket Mode client reconnection - Thanks @seratch @matthieucan
- #1091 #1055 Socket Mode: ENOTSOCK every 300 minutes on Windows - Thanks @dubois @seratch
- #1077 Fix the wrong column data size for bot_token in the built-in SQLAlchemy data model - Thanks @geeorgey
- #1081 Fix a data deletion bug in AmazonS3InstallationStore - Thanks @seratch
- #1085 Add explicit arguments for files_* methods - Thanks @jaebradley
- #1076 Update example code for rate limiting in documents - Thanks @tjstum
- #1071 #1070 Link to Contributor's Guide from README + expand Maintainer's Guide - Thanks @filmaj
- #1072 Fix WebSocket related unit test failures - Thanks @seratch
- All issues/pull requests: https://github.com/slackapi/python-slack-sdk/milestone/40?closed=1
- All changes: v3.8.0...v3.9.0