Slack Tutorial - Slash Commands


This tutorial demonstrates using Cloud Run functions to implement a Slack Slash Command that searches the Google Knowledge Graph API.

Objectives

  • Create a Slash Command in Slack.
  • Write and deploy an HTTP Cloud Run function.
  • Search the Google Knowledge Graph API using the Slash Command.

Costs

In this document, you use the following billable components of Google Cloud:

  • Cloud Run functions
  • Cloud Build
  • Artifact Registry
  • Cloud Logging

For details, see Cloud Run functions pricing.

To generate a cost estimate based on your projected usage, use the pricing calculator. New Google Cloud users might be eligible for a free trial.

Before you begin

  1. Sign in to your Google Cloud account. If you're new to Google Cloud, create an account to evaluate how our products perform in real-world scenarios. New customers also get $300 in free credits to run, test, and deploy workloads.
  2. In the Google Cloud console, on the project selector page, select or create a Google Cloud project.

    Go to project selector

  3. Make sure that billing is enabled for your Google Cloud project.

  4. Enable the Cloud Functions, Cloud Build, Artifact Registry, and Logging APIs.

    Enable the APIs

  5. Install the Google Cloud CLI.
  6. To initialize the gcloud CLI, run the following command:

    gcloud init
  7. In the Google Cloud console, on the project selector page, select or create a Google Cloud project.

    Go to project selector

  8. Make sure that billing is enabled for your Google Cloud project.

  9. Enable the Cloud Functions, Cloud Build, Artifact Registry, and Logging APIs.

    Enable the APIs

  10. Install the Google Cloud CLI.
  11. To initialize the gcloud CLI, run the following command:

    gcloud init
  12. If you already have the gcloud CLI installed, update it by running the following command:

    gcloud components update
  13. Prepare your development environment.

Visualizing the flow of data

The flow of data in the Slack Slash Command tutorial application involves several steps:

  1. The user executes the /kg <search_query> Slash Command in a Slack channel.
  2. Slack sends the command payload to the Cloud Run function's trigger endpoint.
  3. The Cloud Run function sends a request with the user's search query to the Knowledge Graph API.
  4. The Knowledge Graph API responds with any matching results.
  5. The Cloud Run function formats the response into a Slack message.
  6. The Cloud Run function sends the message back to Slack.
  7. The user sees the formatted response in the Slack channel.

It may help to visualize the steps:

Getting credentials

To deploy your function, you need a Google Cloud console-supplied API key and a Slack signing secret.

Getting the Knowledge Graph API key

On the Google Cloud console Credentials page, click the Create credentials button and select API key. Remember this key, as you will include it as part of your deploy command. This key is what enables your function to access the Knowledge Graph API.

Getting the Slack signing secret

You also need the Slack signing secret to deploy your function. To get the Slack signing secret, create the Slack App that will host your Slack Slash Command. This app needs to be associated with a Slack team where you have permissions to install integrations.

  1. Go to the Slack Your Apps page and click Create New App.

  2. Choose From Scratch.

  3. Supply a name for your app and select a Slack workspace where you have permissions to install integrations.

  4. Click Create App.

    The app is created, and the display changes to the Basic Information page.

  5. On the Basic Information page, copy your Slack signing secret and save it.

  6. Save your changes.

Next you need to get the source code and deploy your function. After you deploy your function, you will configure your Slack app to integrate it with the deployed function, as described in Configuring the application.

Preparing the function

  1. Clone the sample app repository to your local machine:

    Node.js

    git clone https://github.com/GoogleCloudPlatform/nodejs-docs-samples.git

    Alternatively, you can download the sample as a zip file and extract it.

    Python

    git clone https://github.com/GoogleCloudPlatform/python-docs-samples.git

    Alternatively, you can download the sample as a zip file and extract it.

    Go

    git clone https://github.com/GoogleCloudPlatform/golang-samples.git

    Alternatively, you can download the sample as a zip file and extract it.

    Java

    git clone https://github.com/GoogleCloudPlatform/java-docs-samples.git

    Alternatively, you can download the sample as a zip file and extract it.

    C#

    git clone https://github.com/GoogleCloudPlatform/dotnet-docs-samples.git

    Alternatively, you can download the sample as a zip file and extract it.

    Ruby

    git clone https://github.com/GoogleCloudPlatform/ruby-docs-samples.git

    Alternatively, you can download the sample as a zip file and extract it.

    PHP

    git clone https://github.com/GoogleCloudPlatform/php-docs-samples.git

    Alternatively, you can download the sample as a zip file and extract it.

  2. Change to the directory that contains the Cloud Run functions sample code:

    Node.js

    cd nodejs-docs-samples/functions/slack/

    Python

    cd python-docs-samples/functions/slack/

    Go

    cd golang-samples/functions/functionsv2/slack/

    Java

    cd java-docs-samples/functions/slack/

    C#

    cd dotnet-docs-samples/functions/slack/SlackKnowledgeGraphSearch/

    Ruby

    cd ruby-docs-samples/functions/slack/

    PHP

    cd php-docs-samples/functions/slack_slash_command/

Deploying the function

To deploy the function that is executed when you (or Slack) make an HTTP POST request to the function's endpoint, run the following command in the directory that contains the sample code (or the pom.xml file for Java).

Replace YOUR_SLACK_SIGNING_SECRET with the signing secret provided by Slack in the Basic information page of your app configuration, and YOUR_KG_API_KEY with the Knowledge Graph API Key you created previously.

Node.js

gcloud functions deploy nodejs-slack-function \
--gen2 \
--runtime=nodejs22 \
--region=REGION \
--source=. \
--entry-point=kgSearch \
--trigger-http \
--set-env-vars "SLACK_SECRET=YOUR_SLACK_SIGNING_SECRET,KG_API_KEY=YOUR_KG_API_KEY" \
--allow-unauthenticated

Use the --runtime flag to specify the runtime ID of a supported Node.js version to run your function.

Python

gcloud functions deploy python-slack-function \
--gen2 \
--runtime=python312 \
--region=REGION \
--source=. \
--entry-point=kg_search \
--trigger-http \
--set-env-vars "SLACK_SECRET=YOUR_SLACK_SIGNING_SECRET,KG_API_KEY=YOUR_KG_API_KEY" \
--allow-unauthenticated

Use the --runtime flag to specify the runtime ID of a supported Python version to run your function.

Go

gcloud functions deploy go-slack-function \
--gen2 \
--runtime=go122 \
--region=REGION \
--source=. \
--entry-point=KGSearch \
--trigger-http \
--set-env-vars "SLACK_SECRET=YOUR_SLACK_SIGNING_SECRET,KG_API_KEY=YOUR_KG_API_KEY" \
--allow-unauthenticated

Use the --runtime flag to specify the runtime ID of a supported Go version to run your function.

Java

gcloud functions deploy java-slack-function \
--gen2 \
--runtime=java21 \
--region=REGION \
--source=. \
--entry-point=functions.SlackSlashCommand \
--memory=512MB \
--trigger-http \
--set-env-vars "SLACK_SECRET=YOUR_SLACK_SIGNING_SECRET,KG_API_KEY=YOUR_KG_API_KEY" \
--allow-unauthenticated

Use the --runtime flag to specify the runtime ID of a supported Java version to run your function.

C#

gcloud functions deploy csharp-slack-function \
--gen2 \
--runtime=dotnet8 \
--region=REGION \
--source=. \
--entry-point=SlackKnowledgeGraphSearch.Function \
--trigger-http \
--set-env-vars "SLACK_SECRET=YOUR_SLACK_SIGNING_SECRET,KG_API_KEY=YOUR_KG_API_KEY" \
--allow-unauthenticated

Use the --runtime flag to specify the runtime ID of a supported .NET version to run your function.

Ruby

gcloud functions deploy ruby-slack-function \
--gen2 \
--runtime=ruby33 \
--region=REGION \
--source=. \
--entry-point=kg_search \
--trigger-http \
--set-env-vars "SLACK_SECRET=YOUR_SLACK_SIGNING_SECRET,KG_API_KEY=YOUR_KG_API_KEY" \
--allow-unauthenticated

Use the --runtime flag to specify the runtime ID of a supported Ruby version to run your function.

PHP

gcloud functions deploy php-slack-function \
--gen2 \
--runtime=php83 \
--region=REGION \
--source=. \
--entry-point=receiveRequest \
--trigger-http \
--set-env-vars "SLACK_SECRET=YOUR_SLACK_SIGNING_SECRET,KG_API_KEY=YOUR_KG_API_KEY" \
--allow-unauthenticated

Use the --runtime flag to specify the runtime ID of a supported PHP version to run your function.

Configuring the application

After the function is deployed, you need to create a Slack Slash Command that sends the query to your function every time the command is triggered:

  1. Go back to the Slack app you created above.

  2. Select the app, go to Slash commands, and click the Create New Command button.

  3. Enter /kg as the name of the command.

  4. In the Request URL field, enter the URL of your function.

    You can either copy the URL from the deploy command output, or go to the Cloud Run functions Overview page in Google Cloud console, click the function to open its Function Details page, and copy the URL from there.

  5. Enter a short description and click Save.

  6. Go to Basic Information.

  7. Click Install to Workspace and follow the instructions on screen to enable the application for your workspace.

    Your Slack Slash Command should come online shortly.

Understanding the code

Importing dependencies

The application must import several dependencies in order to communicate with Google Cloud Platform services:

Node.js

const functions = require('@google-cloud/functions-framework');
const google = require('@googleapis/kgsearch');
const {verifyRequestSignature} = require('@slack/events-api');

// Get a reference to the Knowledge Graph Search component
const kgsearch = google.kgsearch('v1');

Python

import os

from flask import jsonify
import functions_framework
import googleapiclient.discovery
from slack.signature import SignatureVerifier


kgsearch = googleapiclient.discovery.build(
    "kgsearch", "v1", developerKey=os.environ["KG_API_KEY"], cache_discovery=False
)

Go


package slack

import (
	"context"
	"log"
	"os"

	"google.golang.org/api/kgsearch/v1"
	"google.golang.org/api/option"
)

var (
	entitiesService *kgsearch.EntitiesService
	kgKey           string
	slackSecret     string
)

func setup(ctx context.Context) {
	kgKey = os.Getenv("KG_API_KEY")
	slackSecret = os.Getenv("SLACK_SECRET")

	if entitiesService == nil {
		kgService, err := kgsearch.NewService(ctx, option.WithAPIKey(kgKey))
		if err != nil {
			log.Fatalf("kgsearch.NewService: %v", err)
		}
		entitiesService = kgsearch.NewEntitiesService(kgService)
	}
}

Java

private static final Logger logger = Logger.getLogger(SlackSlashCommand.class.getName());
private static final String API_KEY = getenv("KG_API_KEY");
private static final String SLACK_SECRET = getenv("SLACK_SECRET");
private static final Gson gson = new Gson();

private final String apiKey;
private final Kgsearch kgClient;
private final SlackSignature.Verifier verifier;

public SlackSlashCommand() throws IOException, GeneralSecurityException {
  this(new SlackSignature.Verifier(new SlackSignature.Generator(SLACK_SECRET)));
}

SlackSlashCommand(SlackSignature.Verifier verifier) throws IOException, GeneralSecurityException {
  this(verifier, API_KEY);
}

SlackSlashCommand(SlackSignature.Verifier verifier, String apiKey)
    throws IOException, GeneralSecurityException {
  this.verifier = verifier;
  this.apiKey = apiKey;
  this.kgClient = new Kgsearch.Builder(
      GoogleNetHttpTransport.newTrustedTransport(), new GsonFactory(), null).build();
}

// Avoid ungraceful deployment failures due to unset environment variables.
// If you get this warning you should redeploy with the variable set.
private static String getenv(String name) {
  String value = System.getenv(name);
  if (value == null) {
    logger.warning("Environment variable " + name + " was not set");
    value = "MISSING";
  }
  return value;
}

C#

using Google.Apis.Kgsearch.v1;
using Google.Apis.Services;
using Google.Cloud.Functions.Hosting;
using Microsoft.AspNetCore.Hosting;
using Microsoft.Extensions.DependencyInjection;
using System;

// Specify a startup class to use for dependency injection.
// This can also be specified on the function type.
[assembly: FunctionsStartup(typeof(SlackKnowledgeGraphSearch.Startup))]

namespace SlackKnowledgeGraphSearch;

public class Startup : FunctionsStartup
{
    private static readonly TimeSpan SlackTimestampTolerance = TimeSpan.FromMinutes(5);

    public override void ConfigureServices(WebHostBuilderContext context, IServiceCollection services)
    {
        // These can come from an environment variable or a configuration file.
        string kgApiKey = context.Configuration["KG_API_KEY"];
        string slackSigningSecret = context.Configuration["SLACK_SECRET"];

        services.AddSingleton(new KgsearchService(new BaseClientService.Initializer
        {
            ApiKey = kgApiKey
        }));
        services.AddSingleton(new SlackRequestVerifier(slackSigningSecret, SlackTimestampTolerance));
        base.ConfigureServices(context, services);
    }
}

Ruby

require "functions_framework"
require "slack-ruby-client"
require "google/apis/kgsearch_v1"

# This block is executed during cold start, before the function begins
# handling requests. This is the recommended way to create shared resources
# and objects.
FunctionsFramework.on_startup do
  # Create a global handler object, configured with the environment-provided
  # API key and signing secret.
  kg_search = KGSearch.new kg_api_key:     ENV["KG_API_KEY"],
                           signing_secret: ENV["SLACK_SECRET"]
  set_global :kg_search, kg_search
end

# The KGSearch class implements the logic of validating and responding
# to requests. More methods of this class are shown below.
class KGSearch
  def initialize kg_api_key:, signing_secret:
    # Create the global client for the Knowledge Graph Search Service,
    # configuring it with your API key.
    @client = Google::Apis::KgsearchV1::KgsearchService.new
    @client.key = kg_api_key

    # Save signing secret for use by the signature validation method.
    @signing_secret = signing_secret
  end

PHP

use Psr\Http\Message\ServerRequestInterface;
use Psr\Http\Message\ResponseInterface;
use GuzzleHttp\Psr7\Response;

Receiving the webhook

The following function is executed when you (or Slack) make an HTTP POST request to the function's endpoint:

Node.js

/**
 * Receive a Slash Command request from Slack.
 *
 * Trigger this function by creating a Slack slash command with the HTTP Trigger URL.
 * You can find the HTTP URL in the Cloud Console or using `gcloud functions describe`
 *
 * @param {object} req Cloud Function request object.
 * @param {object} req.body The request payload.
 * @param {string} req.rawBody Raw request payload used to validate Slack's message signature.
 * @param {string} req.body.text The user's search query.
 * @param {object} res Cloud Function response object.
 */
functions.http('kgSearch', async (req, res) => {
  try {
    if (req.method !== 'POST') {
      const error = new Error('Only POST requests are accepted');
      error.code = 405;
      throw error;
    }

    if (!req.body.text) {
      const error = new Error('No text found in body.');
      error.code = 400;
      throw error;
    }

    // Verify that this request came from Slack
    verifyWebhook(req);

    // Make the request to the Knowledge Graph Search API
    const response = await makeSearchRequest(req.body.text);

    // Send the formatted message back to Slack
    res.json(response);

    return Promise.resolve();
  } catch (err) {
    console.error(err);
    res.status(err.code || 500).send(err);
    return Promise.reject(err);
  }
});

Python

@functions_framework.http
def kg_search(request):
    if request.method != "POST":
        return "Only POST requests are accepted", 405

    verify_signature(request)
    kg_search_response = make_search_request(request.form["text"])
    return jsonify(kg_search_response)

Go


// Package slack is a Cloud Function which receives a query from
// a Slack command and responds with the KG API result.
package slack

import (
	"context"
	"crypto/hmac"
	"crypto/sha256"
	"encoding/hex"
	"encoding/json"
	"fmt"
	"io"
	"log"
	"net/http"
	"net/url"
	"strconv"
	"strings"
	"time"

	"github.com/GoogleCloudPlatform/functions-framework-go/functions"
)

const (
	version                     = "v0"
	slackRequestTimestampHeader = "X-Slack-Request-Timestamp"
	slackSignatureHeader        = "X-Slack-Signature"
)

type Attachment struct {
	Color     string `json:"color"`
	Title     string `json:"title"`
	TitleLink string `json:"title_link"`
	Text      string `json:"text"`
	ImageURL  string `json:"image_url"`
}

// Message is the a Slack message event.
// see https://api.slack.com/docs/message-formatting
type Message struct {
	ResponseType string       `json:"response_type"`
	Text         string       `json:"text"`
	Attachments  []Attachment `json:"attachments"`
}

func init() {
	functions.HTTP("KGSearch", kgSearch)
	setup(context.Background())
}

// kgSearch uses the Knowledge Graph API to search for a query provided
// by a Slack command.
func kgSearch(w http.ResponseWriter, r *http.Request) {

	bodyBytes, err := io.ReadAll(r.Body)
	if err != nil {
		log.Printf("failed to read body: %v", err)
		http.Error(w, "could not read request body", http.StatusBadRequest)
		return
	}
	if r.Method != "POST" {
		http.Error(w, "Only POST requests are accepted", http.StatusMethodNotAllowed)
		return
	}
	formData, err := url.ParseQuery(string(bodyBytes))
	if err != nil {
		log.Printf("Error: Failed to Parse Form: %v", err)
		http.Error(w, "Couldn't parse form", http.StatusBadRequest)
		return
	}

	result, err := verifyWebHook(r, bodyBytes, slackSecret)
	if err != nil || !result {
		log.Printf("verifyWebhook failed: %v", err)
		http.Error(w, "Failed to verify request signature", http.StatusBadRequest)
		return
	}

	if len(formData.Get("text")) == 0 {
		log.Printf("no search text found: %v", formData)
		http.Error(w, "search text was empty", http.StatusBadRequest)
		return
	}
	kgSearchResponse, err := makeSearchRequest(formData.Get("text"))
	if err != nil {
		log.Printf("makeSearchRequest failed: %v", err)
		http.Error(w, "makeSearchRequest failed", http.StatusInternalServerError)
		return
	}
	w.Header().Set("Content-Type", "application/json")
	if err = json.NewEncoder(w).Encode(kgSearchResponse); err != nil {
		http.Error(w, fmt.Sprintf("failed to marshal results: %v", err), 500)
		return
	}
}

Java

/**
 * Receive a Slash Command request from Slack.
 *
 * @param request Cloud Function request object.
 * @param response Cloud Function response object.
 * @throws IOException if Knowledge Graph request fails
 */
@Override
public void service(HttpRequest request, HttpResponse response) throws IOException {

  // Validate request
  if (!"POST".equals(request.getMethod())) {
    response.setStatusCode(HttpURLConnection.HTTP_BAD_METHOD);
    return;
  }

  // reader can only be read once per request, so we preserve its contents
  String bodyString = request.getReader().lines().collect(Collectors.joining());

  // Slack sends requests as URL-encoded strings
  //   Java 11 doesn't have a standard library
  //   function for this, so do it manually
  Map<String, String> body = new HashMap<>();
  for (String keyValuePair : bodyString.split("&")) {
    String[] keyAndValue = keyValuePair.split("=");
    if (keyAndValue.length == 2) {
      String key = keyAndValue[0];
      String value = keyAndValue[1];

      body.put(key, value);
    }
  }

  if (body == null || !body.containsKey("text")) {
    response.setStatusCode(HttpURLConnection.HTTP_BAD_REQUEST);
    return;
  }

  if (!isValidSlackWebhook(request, bodyString)) {
    response.setStatusCode(HttpURLConnection.HTTP_UNAUTHORIZED);
    return;
  }

  String query = body.get("text");

  // Call knowledge graph API
  JsonObject kgResponse = searchKnowledgeGraph(query);

  // Format response to Slack
  // See https://api.slack.com/docs/message-formatting
  BufferedWriter writer = response.getWriter();

  writer.write(formatSlackMessage(kgResponse, query));

  response.setContentType("application/json");
}

C#

private readonly ILogger _logger;
private readonly KgsearchService _kgService;
private readonly SlackRequestVerifier _verifier;

public Function(ILogger<Function> logger, KgsearchService kgService, SlackRequestVerifier verifier) =>
    (_logger, _kgService, _verifier) = (logger, kgService, verifier);

public async Task HandleAsync(HttpContext context)
{
    var request = context.Request;
    var response = context.Response;
    var cancellationToken = context.RequestAborted;

    // Validate request
    if (request.Method != "POST")
    {
        _logger.LogWarning("Unexpected request method '{method}'", request.Method);
        response.StatusCode = (int) HttpStatusCode.MethodNotAllowed;
        return;
    }

    if (!request.HasFormContentType)
    {
        _logger.LogWarning("Unexpected content type '{contentType}'", request.ContentType);
        response.StatusCode = (int) HttpStatusCode.BadRequest;
        return;
    }

    // We need to read the request body twice: once to validate the signature,
    // and once to read the form content. We copy it into a memory stream,
    // so that we can rewind it after reading.
    var bodyCopy = new MemoryStream();
    await request.Body.CopyToAsync(bodyCopy, cancellationToken);
    request.Body = bodyCopy;
    bodyCopy.Position = 0;

    if (!_verifier.VerifyRequest(request, bodyCopy.ToArray()))
    {
        _logger.LogWarning("Slack request verification failed");
        response.StatusCode = (int) HttpStatusCode.Unauthorized;
        return;
    }

    var form = await request.ReadFormAsync();
    if (!form.TryGetValue("text", out var query))
    {
        _logger.LogWarning("Slack request form did not contain a text element");
        response.StatusCode = (int) HttpStatusCode.BadRequest;
        return;
    }

    var kgResponse = await SearchKnowledgeGraphAsync(query, cancellationToken);
    string formattedResponse = FormatSlackMessage(kgResponse, query);
    response.ContentType = "application/json";
    await response.WriteAsync(formattedResponse);
}

Ruby

# Handler for the function endpoint.
FunctionsFramework.http "kg_search" do |request|
  # Return early if the request is not a POST.
  unless request.post?
    return [405, {}, ["Only POST requests are accepted."]]
  end

  # Access the global Knowledge Graph Search client
  kg_search = global :kg_search

  # Verify the request signature and return early if it failed.
  unless kg_search.signature_valid? request
    return [401, {}, ["Signature validation failed."]]
  end

  # Query the Knowledge Graph and format a Slack message with the response.
  # This method returns a nested hash, which the Functions Framework will
  # convert to JSON automatically.
  kg_search.make_search_request request.params["text"]
end

PHP

/**
 * Receive a Slash Command request from Slack.
 */
function receiveRequest(ServerRequestInterface $request): ResponseInterface
{
    // Validate request
    if ($request->getMethod() !== 'POST') {
        // [] = empty headers
        return new Response(405);
    }

    // Parse incoming URL-encoded requests from Slack
    // (Slack requests use the "application/x-www-form-urlencoded" format)
    $bodyStr = $request->getBody();
    parse_str($bodyStr, $bodyParams);

    if (!isset($bodyParams['text'])) {
        // [] = empty headers
        return new Response(400);
    }

    if (!isValidSlackWebhook($request, $bodyStr)) {
        // [] = empty headers
        return new Response(403);
    }

    $query = $bodyParams['text'];

    // Call knowledge graph API
    $kgResponse = searchKnowledgeGraph($query);

    // Format response to Slack
    // See https://api.slack.com/docs/message-formatting
    $formatted_message = formatSlackMessage($kgResponse, $query);

    return new Response(
        200,
        ['Content-Type' => 'application/json'],
        $formatted_message
    );
}

The following function authenticates the incoming request by verifying the X-Slack-Signature header provided by Slack:

Node.js

/**
 * Verify that the webhook request came from Slack.
 *
 * @param {object} req Cloud Function request object.
 * @param {string} req.headers Headers Slack SDK uses to authenticate request.
 * @param {string} req.rawBody Raw body of webhook request to check signature against.
 */
const verifyWebhook = req => {
  const signature = {
    signingSecret: process.env.SLACK_SECRET,
    requestSignature: req.headers['x-slack-signature'],
    requestTimestamp: req.headers['x-slack-request-timestamp'],
    body: req.rawBody,
  };

  // This method throws an exception if an incoming request is invalid.
  verifyRequestSignature(signature);
};

Python

def verify_signature(request):
    request.get_data()  # Decodes received requests into request.data

    verifier = SignatureVerifier(os.environ["SLACK_SECRET"])

    if not verifier.is_valid_request(request.data, request.headers):
        raise ValueError("Invalid request/credentials.")

Go


// verifyWebHook verifies the request signature.
// See https://api.slack.com/docs/verifying-requests-from-slack.
func verifyWebHook(r *http.Request, body []byte, slackSigningSecret string) (bool, error) {
	timeStamp := r.Header.Get(slackRequestTimestampHeader)
	slackSignature := r.Header.Get(slackSignatureHeader)

	t, err := strconv.ParseInt(timeStamp, 10, 64)
	if err != nil {
		return false, fmt.Errorf("strconv.ParseInt(%s): %w", timeStamp, err)
	}

	if ageOk, age := checkTimestamp(t); !ageOk {
		return false, fmt.Errorf("checkTimestamp(%v): %v %v", t, ageOk, age)
	}

	if timeStamp == "" || slackSignature == "" {
		return false, fmt.Errorf("timeStamp and/or signature headers were blank")
	}

	baseString := fmt.Sprintf("%s:%s:%s", version, timeStamp, body)

	signature := getSignature([]byte(baseString), []byte(slackSigningSecret))

	trimmed := strings.TrimPrefix(slackSignature, fmt.Sprintf("%s=", version))
	signatureInHeader, err := hex.DecodeString(trimmed)

	if err != nil {
		return false, fmt.Errorf("hex.DecodeString(%v): %w", trimmed, err)
	}

	return hmac.Equal(signature, signatureInHeader), nil
}

func getSignature(base []byte, secret []byte) []byte {
	h := hmac.New(sha256.New, secret)
	h.Write(base)

	return h.Sum(nil)
}

// checkTimestamp allows requests time stamped less than 5 minutes ago.
func checkTimestamp(timeStamp int64) (bool, time.Duration) {
	t := time.Since(time.Unix(timeStamp, 0))

	return t.Minutes() <= 5, t
}

Java

/**
 * Verify that the webhook request came from Slack.
 *
 * @param request Cloud Function request object in {@link HttpRequest} format.
 * @param requestBody Raw body of webhook request to check signature against.
 * @return true if the provided request came from Slack, false otherwise
 */
boolean isValidSlackWebhook(HttpRequest request, String requestBody) {
  // Check for headers
  Optional<String> maybeTimestamp = request.getFirstHeader("X-Slack-Request-Timestamp");
  Optional<String> maybeSignature = request.getFirstHeader("X-Slack-Signature");
  if (!maybeTimestamp.isPresent() || !maybeSignature.isPresent()) {
    return false;
  }

  Long nowInMs = ZonedDateTime.now().toInstant().toEpochMilli();

  return verifier.isValid(maybeTimestamp.get(), requestBody, maybeSignature.get(), nowInMs);
}

C#

using Microsoft.AspNetCore.Http;
using System;
using System.Security.Cryptography;
using System.Text;

namespace SlackKnowledgeGraphSearch;

public class SlackRequestVerifier
{
    private const string Version = "v0";
    private const string SignaturePrefix = Version + "=";
    private static readonly byte[] VersionBytes = Encoding.UTF8.GetBytes(Version);
    private static readonly byte[] ColonBytes = Encoding.UTF8.GetBytes(":");

    private readonly byte[] _signingKeyBytes;
    private readonly TimeSpan _timestampTolerance;

    public SlackRequestVerifier(string signingKey, TimeSpan timestampTolerance) =>
        (_signingKeyBytes, _timestampTolerance) =
        (Encoding.UTF8.GetBytes(signingKey), timestampTolerance);

    /// <summary>
    /// Verifies a request signature with the given content, which would
    /// normally have been read from the request beforehand.
    /// </summary>
    public bool VerifyRequest(HttpRequest request, byte[] content)
    {
        if (!request.Headers.TryGetValue("X-Slack-Request-Timestamp", out var timestampHeader) ||
            !request.Headers.TryGetValue("X-Slack-Signature", out var signatureHeader))
        {
            return false;
        }

        // Validate that the request isn't too old.
        if (!int.TryParse(timestampHeader, out var unixSeconds))
        {
            return false;
        }
        var timestamp = DateTimeOffset.FromUnixTimeSeconds(unixSeconds);
        if (timestamp + _timestampTolerance < DateTimeOffset.UtcNow)
        {
            return false;
        }

        // Hash the version:timestamp:content
        using var hmac = new HMACSHA256(_signingKeyBytes);
        AddToHash(VersionBytes);
        AddToHash(ColonBytes);
        AddToHash(Encoding.UTF8.GetBytes(timestampHeader));
        AddToHash(ColonBytes);
        byte[] hash = hmac.ComputeHash(content);
        void AddToHash(byte[] bytes) => hmac.TransformBlock(bytes, 0, bytes.Length, null, 0);

        // Compare the resulting signature with the signature in the request.
        return CreateSignature(hash) == (string) signatureHeader;
    }

    private const string Hex = "0123456789abcdef";
    private static string CreateSignature(byte[] hash)
    {
        var builder = new StringBuilder(SignaturePrefix, SignaturePrefix.Length + hash.Length * 2);
        foreach (var b in hash)
        {
            builder.Append(Hex[b >> 4]);
            builder.Append(Hex[b & 0xf]);
        }
        return builder.ToString();
    }
}

Ruby

# slack-ruby-client expects a Rails-style request object with a "headers"
# method, but the Functions Framework provides only a Rack request.
# To avoid bringing in Rails as a dependency, we'll create a simple class
# that implements the "headers" method and delegates everything else back to
# the Rack request object.
require "delegate"
class RequestWithHeaders < SimpleDelegator
  def headers
    env.each_with_object({}) do |(key, val), result|
      if /^HTTP_(\w+)$/ =~ key
        header = Regexp.last_match(1).split("_").map(&:capitalize).join("-")
        result[header] = val
      end
    end
  end
end

# This is a method of the KGSearch class.
# It determines whether the given request's signature is valid.
def signature_valid? request
  # Wrap the request with our class that provides the "headers" method.
  request = RequestWithHeaders.new request

  # Validate the request signature.
  slack_request = Slack::Events::Request.new request,
                                             signing_secret: @signing_secret
  slack_request.valid?
end

PHP

/**
 * Verify that the webhook request came from Slack.
 */
function isValidSlackWebhook(ServerRequestInterface $request): bool
{
    $SLACK_SECRET = getenv('SLACK_SECRET');

    // Check for headers
    $timestamp = $request->getHeaderLine('X-Slack-Request-Timestamp');
    $signature = $request->getHeaderLine('X-Slack-Signature');
    if (!$timestamp || !$signature) {
        return false;
    }

    // Compute signature
    $plaintext = sprintf('v0:%s:%s', $timestamp, $request->getBody());
    $hash = sprintf('v0=%s', hash_hmac('sha256', $plaintext, $SLACK_SECRET));

    return $hash === $signature;
}

Querying the Knowledge Graph API

The following function sends a request with the user's search query to the Knowledge Graph API:

Node.js

/**
 * Send the user's search query to the Knowledge Graph API.
 *
 * @param {string} query The user's search query.
 */
const makeSearchRequest = query => {
  return new Promise((resolve, reject) => {
    kgsearch.entities.search(
      {
        auth: process.env.KG_API_KEY,
        query: query,
        limit: 1,
      },
      (err, response) => {
        console.log(err);
        if (err) {
          reject(err);
          return;
        }

        // Return a formatted message
        resolve(formatSlackMessage(query, response));
      }
    );
  });
};

Python

def make_search_request(query):
    req = kgsearch.entities().search(query=query, limit=1)
    res = req.execute()
    return format_slack_message(query, res)

Go

func makeSearchRequest(query string) (*Message, error) {
	res, err := entitiesService.Search().Query(query).Limit(1).Do()
	if err != nil {
		return nil, err
	}
	return formatSlackMessage(query, res)
}

Java

/**
 * Send the user's search query to the Knowledge Graph API.
 *
 * @param query The user's search query.
 * @return The Knowledge graph API results as a {@link JsonObject}.
 * @throws IOException if Knowledge Graph request fails
 */
JsonObject searchKnowledgeGraph(String query) throws IOException {
  Kgsearch.Entities.Search kgRequest = kgClient.entities().search();
  kgRequest.setQuery(query);
  kgRequest.setKey(apiKey);

  return gson.fromJson(kgRequest.execute().toString(), JsonObject.class);
}

C#

private async Task<SearchResponse> SearchKnowledgeGraphAsync(string query, CancellationToken cancellationToken)
{
    _logger.LogInformation("Performing Knowledge Graph search for '{query}'", query);
    var request = _kgService.Entities.Search();
    request.Limit = 1;
    request.Query = query;
    return await request.ExecuteAsync(cancellationToken);
}

Ruby

# This is a method of the KGSearch class.
# It makes an API call to the Knowledge Graph Search Service, and formats
# a Slack message as a nested Hash object.
def make_search_request query
  response = @client.search_entities query: query, limit: 1
  format_slack_message query, response
end

PHP

/**
 * Send the user's search query to the Knowledge Graph API.
 */
function searchKnowledgeGraph(string $query): Google_Service_Kgsearch_SearchResponse
{
    $API_KEY = getenv('KG_API_KEY');

    $apiClient = new Google\Client();
    $apiClient->setDeveloperKey($API_KEY);

    $service = new Google_Service_Kgsearch($apiClient);

    $params = ['query' => $query];

    $kgResults = $service->entities->search($params);

    return $kgResults;
}

Formatting the Slack message

Finally, the following function formats the Knowledge Graph result into a richly formatted Slack message that will be displayed to the user:

Node.js

/**
 * Format the Knowledge Graph API response into a richly formatted Slack message.
 *
 * @param {string} query The user's search query.
 * @param {object} response The response from the Knowledge Graph API.
 * @returns {object} The formatted message.
 */
const formatSlackMessage = (query, response) => {
  let entity;

  // Extract the first entity from the result list, if any
  if (
    response &&
    response.data &&
    response.data.itemListElement &&
    response.data.itemListElement.length > 0
  ) {
    entity = response.data.itemListElement[0].result;
  }

  // Prepare a rich Slack message
  // See https://api.slack.com/docs/message-formatting
  const slackMessage = {
    response_type: 'in_channel',
    text: `Query: ${query}`,
    attachments: [],
  };

  if (entity) {
    const attachment = {
      color: '#3367d6',
    };
    if (entity.name) {
      attachment.title = entity.name;
      if (entity.description) {
        attachment.title = `${attachment.title}: ${entity.description}`;
      }
    }
    if (entity.detailedDescription) {
      if (entity.detailedDescription.url) {
        attachment.title_link = entity.detailedDescription.url;
      }
      if (entity.detailedDescription.articleBody) {
        attachment.text = entity.detailedDescription.articleBody;
      }
    }
    if (entity.image && entity.image.contentUrl) {
      attachment.image_url = entity.image.contentUrl;
    }
    slackMessage.attachments.push(attachment);
  } else {
    slackMessage.attachments.push({
      text: 'No results match your query...',
    });
  }

  return slackMessage;
};

Python

def format_slack_message(query, response):
    entity = None
    if (
        response
        and response.get("itemListElement") is not None
        and len(response["itemListElement"]) > 0
    ):
        entity = response["itemListElement"][0]["result"]

    message = {
        "response_type": "in_channel",
        "text": f"Query: {query}",
        "attachments": [],
    }

    attachment = {}
    if entity:
        name = entity.get("name", "")
        description = entity.get("description", "")
        detailed_desc = entity.get("detailedDescription", {})
        url = detailed_desc.get("url")
        article = detailed_desc.get("articleBody")
        image_url = entity.get("image", {}).get("contentUrl")

        attachment["color"] = "#3367d6"
        if name and description:
            attachment["title"] = "{}: {}".format(entity["name"], entity["description"])
        elif name:
            attachment["title"] = name
        if url:
            attachment["title_link"] = url
        if article:
            attachment["text"] = article
        if image_url:
            attachment["image_url"] = image_url
    else:
        attachment["text"] = "No results match your query."
    message["attachments"].append(attachment)

    return message

Go


package slack

import (
	"encoding/json"
	"fmt"

	"google.golang.org/api/kgsearch/v1"
)

type ItemList struct {
	Items []ItemListElement `json:"itemListElement"`
}
type ItemListElement struct {
	Result EntitySearchResult `json:"result"`
}
type EntitySearchResult struct {
	Name         string       `json:"name"`
	Description  string       `json:"description"`
	DetailedDesc DetailedDesc `json:"detailedDescription"`
	URL          string       `json:"url"`
	Image        Image
}
type DetailedDesc struct {
	ArticleBody string
	URL         string
}
type Image struct {
	ContentURL string
}

func formatSlackMessage(query string, response *kgsearch.SearchResponse) (*Message, error) {
	if response == nil {
		return nil, fmt.Errorf("empty response")
	}

	if response.ItemListElement == nil || len(response.ItemListElement) == 0 {
		message := &Message{
			ResponseType: "in_channel",
			Text:         fmt.Sprintf("Query: %s", query),
			Attachments: []Attachment{
				{
					Color: "#d6334b",
					Text:  "No results match your query.",
				},
			},
		}
		return message, nil
	}

	// The KnowledgeGraph API returns an empty interface. To make this more
	// useful, we convert it back to json, and unmarshal into specific types.
	jsonstring, err := response.MarshalJSON()
	if err != nil {
		return nil, err
	}
	r := &ItemList{}
	if err := json.Unmarshal(jsonstring, r); err != nil {
		return nil, fmt.Errorf("failed to unmarshal json: %w", err)
	}
	result := r.Items[0].Result

	attach := Attachment{Color: "#3367d6"}
	attach.Title = result.Name
	attach.TitleLink = result.DetailedDesc.URL
	attach.Text = result.DetailedDesc.ArticleBody
	attach.ImageURL = result.Image.ContentURL

	message := &Message{
		ResponseType: "in_channel",
		Text:         fmt.Sprintf("Query: %s", query),
		Attachments:  []Attachment{attach},
	}
	return message, nil
}

Java

/**
 * Helper method to copy properties between {@link JsonObject}s
 */
void addPropertyIfPresent(
    JsonObject target, String targetName, JsonObject source, String sourceName) {
  if (source.has(sourceName)) {
    target.addProperty(targetName, source.get(sourceName).getAsString());
  }
}

/**
 * Format the Knowledge Graph API response into a richly formatted Slack message.
 *
 * @param kgResponse The response from the Knowledge Graph API as a {@link JsonObject}.
 * @param query The user's search query.
 * @return The formatted Slack message as a JSON string.
 */
String formatSlackMessage(JsonObject kgResponse, String query) {
  JsonObject attachmentJson = new JsonObject();

  JsonObject responseJson = new JsonObject();
  responseJson.addProperty("response_type", "in_channel");
  responseJson.addProperty("text", String.format("Query: %s", query));

  JsonArray entityList = kgResponse.getAsJsonArray("itemListElement");

  // Extract the first entity from the result list, if any
  if (entityList.size() == 0) {
    attachmentJson.addProperty("text", "No results match your query...");
    responseJson.add("attachments", attachmentJson);

    return gson.toJson(responseJson);
  }

  JsonObject entity = entityList.get(0).getAsJsonObject().getAsJsonObject("result");

  // Construct Knowledge Graph response attachment
  String title = entity.get("name").getAsString();
  if (entity.has("description")) {
    title = String.format("%s: %s", title, entity.get("description").getAsString());
  }
  attachmentJson.addProperty("title", title);

  if (entity.has("detailedDescription")) {
    JsonObject detailedDescJson = entity.getAsJsonObject("detailedDescription");
    addPropertyIfPresent(attachmentJson, "title_link", detailedDescJson, "url");
    addPropertyIfPresent(attachmentJson, "text", detailedDescJson, "articleBody");
  }

  if (entity.has("image")) {
    JsonObject imageJson = entity.getAsJsonObject("image");
    addPropertyIfPresent(attachmentJson, "image_url", imageJson, "contentUrl");
  }

  JsonArray attachmentList = new JsonArray();
  attachmentList.add(attachmentJson);

  responseJson.add("attachments", attachmentList);

  return gson.toJson(responseJson);
}

C#

private string FormatSlackMessage(SearchResponse kgResponse, string query)
{
    JObject attachment = new JObject();
    JObject response = new JObject();

    response["response_type"] = "in_channel";
    response["text"] = $"Query: {query}";

    var element = kgResponse.ItemListElement?.FirstOrDefault() as JObject;
    if (element is object && element.TryGetValue("result", out var entityToken) &&
        entityToken is JObject entity)
    {
        string title = (string) entity["name"];
        if (entity.TryGetValue("description", out var description))
        {
            title = $"{title}: {description}";
        }
        attachment["title"] = title;
        if (entity.TryGetValue("detailedDescription", out var detailedDescriptionToken) &&
            detailedDescriptionToken is JObject detailedDescription)
        {
            AddPropertyIfPresent(detailedDescription, "url", "title_link");
            AddPropertyIfPresent(detailedDescription, "articleBody", "text");
        }
        if (entity.TryGetValue("image", out var imageToken) &&
            imageToken is JObject image)
        {
            AddPropertyIfPresent(image, "contentUrl", "image_url");
        }
    }
    else
    {
        attachment["text"] = "No results match your query...";
    }
    response["attachments"] = new JArray { attachment };
    return response.ToString();

    void AddPropertyIfPresent(JObject parent, string sourceProperty, string targetProperty)
    {
        if (parent.TryGetValue(sourceProperty, out var propertyValue))
        {
            attachment[targetProperty] = propertyValue;
        }
    }
}

Ruby

# This is a method of the KGSearch class.
# It takes a raw SearchResponse from the Knowledge Graph Search Service,
# and formats a Slack message.
def format_slack_message query, response
  result = response.item_list_element&.first&.fetch "result", nil
  attachment =
    if result
      name = result.fetch "name", nil
      description = result.fetch "description", nil
      details = result.fetch "detailedDescription", {}
      { "title"      => name && description ? "#{name}: #{description}" : name,
        "title_link" => details.fetch("url", nil),
        "text"       => details.fetch("articleBody", nil),
        "image_url"  => result.fetch("image", nil)&.fetch("contentUrl", nil) }
    else
      { "text" => "No results match your query." }
    end
  { "response_type" => "in_channel",
    "text"          => "Query: #{query}",
    "attachments"   => [attachment.compact] }
end

PHP

/**
 * Format the Knowledge Graph API response into a richly formatted Slack message.
 */
function formatSlackMessage(Google_Service_Kgsearch_SearchResponse $kgResponse, string $query): string
{
    $responseJson = [
        'response_type' => 'in_channel',
        'text' => 'Query: ' . $query
    ];

    $entityList = $kgResponse['itemListElement'];

    // Extract the first entity from the result list, if any
    if (empty($entityList)) {
        $attachmentJson = ['text' => 'No results match your query...'];
        $responseJson['attachments'] = $attachmentJson;

        return json_encode($responseJson);
    }

    $entity = $entityList[0]['result'];

    // Construct Knowledge Graph response attachment
    $title = $entity['name'];
    if (isset($entity['description'])) {
        $title = $title . ' ' . $entity['description'];
    }
    $attachmentJson = ['title' => $title];

    if (isset($entity['detailedDescription'])) {
        $detailedDescJson = $entity['detailedDescription'];
        $attachmentJson = array_merge([
            'title_link' => $detailedDescJson[ 'url'],
            'text' => $detailedDescJson['articleBody'],
        ], $attachmentJson);
    }

    if (isset($entity['image'])) {
        $imageJson = $entity['image'];
        $attachmentJson['image_url'] = $imageJson['contentUrl'];
    }

    $responseJson['attachments'] = array($attachmentJson);

    return json_encode($responseJson);
}

Slack API timeouts

The Slack API expects your function to respond within 3 seconds of receiving a webhook request.

The commands in this tutorial typically take less than 3 seconds to respond. For longer-running commands, we recommend configuring a function to push requests (including their response_url field) to a Pub/Sub topic that acts as a task queue.

Then, you can create a second function triggered by Pub/Sub that processes those tasks and sends results back to Slack's response_url.

Using the Slash command

  1. Once the function deployment has completed, type this command into your Slack channel:

    /kg giraffe

    You should see the Knowledge Graph entry for "giraffe."

  2. Check the logs to see the output from the function execution:

    gcloud functions logs read --limit 100
    

Clean up

To avoid incurring charges to your Google Cloud account for the resources used in this tutorial, either delete the project that contains the resources, or keep the project and delete the individual resources.

Deleting the project

The easiest way to eliminate billing is to delete the project that you created for the tutorial.

To delete the project:

  1. In the Google Cloud console, go to the Manage resources page.

    Go to Manage resources

  2. In the project list, select the project that you want to delete, and then click Delete.
  3. In the dialog, type the project ID, and then click Shut down to delete the project.

Deleting the function

To delete the Cloud Run function you deployed in this tutorial, run the following command:

Node.js

gcloud functions delete nodejs-slack-function --gen2 --region REGION 

Python

gcloud functions delete python-slack-function --gen2 --region REGION 

Go

gcloud functions delete go-slack-function --gen2 --region REGION 

Java

gcloud functions delete java-slack-function --gen2 --region REGION 

C#

gcloud functions delete csharp-slack-function --gen2 --region REGION 

Ruby

gcloud functions delete ruby-slack-function --gen2 --region REGION 

PHP

gcloud functions delete php-slack-function --gen2 --region REGION 

You can also delete Cloud Run functions from the Google Cloud console.