Skip to content

Deploying a platform with Docker

In the previous sections to discover and test Tock, you used the demo platform. This allowed you to discover the construction and configuration of Tock bots without having to install the platform first.

In this guide, you will learn how to deploy a complete Tock platform in a few minutes, thanks to the Docker / Docker Compose implementation examples provided.

Note that it is entirely possible to deploy Tock without using Docker.

What you will create

  • A complete local Tock platform: Tock Studio, Bot API, etc.

  • A bot and a minimal configuration to test the platform

  • (Optional) A Kotlin program connecting to the local platform via WebSocket

Prerequisites

  • About 20 minutes

  • To deploy the platform locally, a development environment with recent versions of Docker and Docker Compose installed

If you don't want to use Docker, no problem. There are other ways to deploy the MongoDB database and Kotlin services on JVM. However, you can browse the Dockerfile and docker-compose.yml as examples to instantiate these services.

  • (Optional) For the WebSocket program, a development environment (or IDE) supporting Kotlin, for example IntelliJ with recent versions of JDK and Maven

Without IDE or without Maven, no problem. It is quite possible to compile and run the program with other tools.

Deploy a Tock platform - without sources

It is possible to retrieve only a few files from the GitHub repository, without downloading all the Tock sources. In a few command lines, the platform is operational.

However, it is essential to have recent versions of Docker and Docker Compose.

To start from the sources of the Tock Docker repository, go instead to the next paragraph. 

# Get the lastest docker-compose from GitHub (including Bot API)
$ curl -o docker-compose.yml https://raw.githubusercontent.com/theopenconversationkit/tock-docker/master/docker-compose-bot.yml
# Get the lastest database-init script from GitHub
$ mkdir -p scripts && curl -o scripts/setup.sh https://raw.githubusercontent.com/theopenconversationkit/tock-docker/master/scripts/setup.sh && chmod +x scripts/setup.sh
# Get the lastest Tock version/tag from GitHub
$ curl -o .env https://raw.githubusercontent.com/theopenconversationkit/tock-docker/master/.env

# Run the stack
$ docker-compose up

Deploying a Tock platform - from sources

This is an alternative way to start Tock, from the Tock Docker repository.

In addition to Docker and Docker Compose, you need either a Git client to retrieve the sources (git clone command) or to have already copied the GitHub sources locally.

To start without Git or the repository sources locally, follow the previous paragraph. 

# Get the lastest sources from GitHub
$ git clone https://github.com/theopenconversationkit/tock-docker.git && cd tock-docker
# Make the database-init script executable
$ chmod +x scripts/setup.sh

# Run the stack (including Bot API)
$ docker-compose -f docker-compose-bot.yml up

Accessing Tock Studio

Once the platform is ready, the Tock Studio interfaces are on port 80 by default:

After deploying the platform, it initializes, and you may have to wait a few seconds before the Tock Studio interfaces are accessible.

  • Log in with the default admin@app.com / password credentials

It is obviously recommended to change these values ​​when installing a platform intended for long-term use (production, platform shared between teams, etc.).

Create an application, a connector and an intention

As in the first bot guide using the demo platform, you will create a Tock application and a connector to start using the local platform. Feel free to go back to the previous guides for more comments.

When you first access the local platform:

  • Enter a name for the application

  • Select a language - you can add others later

  • Validate to create the application

  • Go to Settings > Configurations

  • Create a new Configuration

  • Select the Slack connector type

  • Create

Note the API Key automatically generated for your application. It will be useful if you try the WebSocket mode later in this guide (optional).

  • Go to Stories & Answers > New Story

  • Enter a user phrase for example "hello"

  • In the Add new Answer field, enter a response for example "what a beautiful day!"

  • Finish with Create Story

  • Go to Test > Test the Bot

  • Say "hello" 🙋, the bot answers you 🤖

Connect a course in Kotlin (optional)

As in the guide program courses using the demo platform, you will create a Kotlin application connecting in WebSocket to the local Tock platform. Feel free to go back to the previous guides for more details.

  • Create a Kotlin project for example with Maven as indicated in the guide program courses

The classpath must include tock-bot-api-websocket to use the WebSocket mode.

  • Create a Kotlin file (e.g. in `src/main/kotlin/StartWebSocket.kt)

  • Edit it with the following code:

import ai.tock.bot.api.client.newBot
import ai.tock.bot.api.client.newStory
import ai.tock.bot.api.websocket.start

fun main() {
    start( // Do not use #startWithDemo when integrating with a local platform 
        newBot(
            "PUT-YOUR-TOCK-APP-API-KEY-HERE", // Get your app API key from Bot Configurations in Tock Studio
            newStory("qui-es-tu") { // Answer for the 'qui-es-tu' story
                send("Je suis un assistant conversationnel construit avec Tock")
                end("Comment puis-je aider ?")
            }
        ),
        "http://localhost:8080" // Local platform URL (default host/port)
    ) 
}

You can find this code (and other examples) in the tock-bot-samples repository.

  • Replace the API key with the one of your own Tock application. To do this, in Tock Studio, go to Settings > Configurations and report the API Key value in the code.

  • Run the function (main) in your development environment.

  • Go back to Tock in Test > Test the Bot and say "who are you?": the bot does not answer yet.

  • Go to Language Understanding > Inbox, for the sentence you just entered:

  • Change the intent to New intent

  • Name it "who-are-you" as in the code (so that the link is made)

  • Create the intent with Create

  • Finish qualifying the sentence with Validate

  • Go back to Test > Test the Bot. Say "who are you?": the bot answers!

Congratulations!

You have just deployed your own Tock conversational platform locally.

This can be used to better understand the architecture and check the portability of the solution, but also during developments, for Tock contributors or if you have to work without Internet access (on the move, on a restricted network, etc.).

Warning, the provided Docker implementation is not enough to guarantee resilience and scalability of the platform whatever the production conditions. For this, some recommendations are proposed in the high availability section of the Tock manual.

Continue...

You have just completed the Tock quick start guides.

From here, you can launch directly on a Tock platform.

Other pages also present customer case studies, code examples, how to contact the Tock community, etc.

Chat with Tock
×