Skip to content

Latest commit

 

History

History
53 lines (28 loc) · 2.23 KB

README.md

File metadata and controls

53 lines (28 loc) · 2.23 KB
Raw

TagFinderBot

Wanna see working bot 🤖

Check - https://t.me/TagFinderBot

This is a tutorial to create a telegram bot which can suggest tags/labels in an image using AWS Rekognition. We will be using TelegramBots Java API.

Steps to create the bot -

  1. Create new bot and get token from https://t.me/BotFather by clicking on /newbot command. Set Description, About, Botpic if you want to.

  2. Create a java maven project

  3. Add TelegramBots API dependency.

    <dependency>
    <groupId>org.telegram</groupId>
    <artifactId>telegrambots</artifactId>
    <version>4.1</version>
</dependency>

and AWS Java SDK Rekognition dependency.

   <dependency>
       <groupId>com.amazonaws</groupId>
       <artifactId>aws-java-sdk-rekognition</artifactId>
       <version>1.11.368</version>
   </dependency>

Note: Make sure to use updated dependencies.

  1. Create a new bot by extending org.telegram.telegrambots.bots.TelegramLongPollingBot. Set Bot username in getBotUsername() and Bot token in getBotToken(). We will recieve updates in onUpdateReceived() method.

  2. Now create logic to get the picture sent by the user and get ByteBuffer from it which can be fed to AWS Rekognition to find tags. Check TagFinderBot.java for details.

  3. Get ACCESS_KEY and SECRET_KEY from AWS to access AWS Rekognition. Check AWSRekognition.java for code to get tags/labels from the ByteBuffer.

  4. Now finally register your bot. Check TagFinderApp.java for details.

  5. All done. Build the project mvn clean install. tagfinder-0.1-jar-with-dependencies.jar will be created in target folder.

  6. Start your bot 🤖 using

    java -jar tagfinder-0.1-jar-with-dependencies.jar