Bilbot

Design of an extensible chatbots cloud architecture for its application to the improvement of the consumption and the exploitation of public services and data.

Screenshots

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

• Docker
• Docker compose

How to run

This project is fully Dockerized and orchestrated with Docker Compose, is mandatory to have previously installed both.

Create your version of src/bilbot-php/www/Constants.php.dist without the .dist extension and your API keys.

Build the images:

docker-compose build

Run the services:

docker-compose up -d

Prepare the environment:

1. Log in to the watson-php container by running the following command:

   docker exec -it bilbotpfm_watson-php_1 bash
   

2. Install dependencies by running the following command:

   cd /var/www/docker-symfony && composer install -n
   

3. Change /tmp dir permissions:

   chown -R www-data:www-data /tmp
   

Do the same process with welive-php container.

Open Telegram and search the bot @bilb0_bot to start a conversation.

Commands

You can ask to the bot information using the following commands:

• /saludo for a greeting message.
• /agenda for information about events.
• /bicis for information about bike rental service.
• /asociaciones for information about clubs.
• /hoteles for information about accomodation.
• /restaurantes for information about restaurants.
• /atracciones for information about tourist attractions.
• /turismo for information about tourist offices.

Get fun with it asking information!

If you want to access directly to microservices from your computer update your hosts

Mac OS X

Add the following line to your /etc/hosts file:

127.0.0.1 bilbot.dev

Windows

Add the following line to your %SystemRoot%\System32\drivers\etc\hosts file:

127.0.0.1 bilbot.dev

The web server is constantly querying the Telegram API for new messages sent to Bilbot, for stopping the containers:

docker-compose down

To remove the built image and starting from scratch:

docker-compose down
docker-compose rm

All the received messages are stored in a MySQL Database which can be accessed in localhost.

Microservices

Name	Address	Description
bilbot	bilbot.dev:80	Contains the Telegram bot API logic and the user commands.
bilbot-watson	bilbot.dev:81	Contains the Watson API querying logic and is based on Symfony. Contains a testing index page.
bilbot-welive	bilbot.dev:82	Contains the WeLive API querying logic and is based on Symfony. Contains a testing index page.

Deployment

The deployment process is based on a Bash script, which compress the source folders in a tar and upload them to a VPS.

For deploying a new version in your VPS, just add your SSH keys to your keychain, change the VPS address in deploy.sh and execute:

./deploy.sh

If you want to stop the bot instance in your VPS, change the VPS address of the file stop_vps.sh and execute it with

./stop_vps.sh

How to add a new command

1. Create a new Controller with all the needed queries to the dataset inside of src/bilbot-welive/src/AppBundle/Controller/REST
   1. Create at least 3 types of queries listAction, detailAction and searchAction for making easier the interaction with the bot
2. Add the routing for the actions inside of src/bilbot-welive/src/AppBundle/Resources/config/routing_api.yml following the naming convention {dataType}_{action_name}
3. Create a new command file in src/bilbot-php/www/Commands with the name {CommandNameCommand.php} is highly recommended to duplicate an existing one and changing the name of the class within
4. Change the class properties $name, $description, $usage and the constants WELIVE_SEARCH_METHOD, WELIVE_LIST_METHOD and DATA_PREFIX, please note that these are the route slugs created in step 2. DATA_PREFIX is recommended to be one word with a trailing underscore (_) should be descriptive of the type of data, this will be used by the CallbackqueryCommand.php for handling the origin of the callbacks.
5. Customize the button text representation in search function call parameters and the data column to be represented as title.
6. Add the data prefix without the trailing underscore to the $entities associative array in CallbackqueryCommand.php with the subkeys action to represent the detail route slug created in step 2 and column for the column that will be used as primary key for the search, it must be consistent with the argument that will receive the detail action in the controller.
7. Add a new switch clause in the function buildAnswer using the data prefix without trailing underscore with the text that will be rendered as answer to the user.
8. Add the new file command path to src/bilbot-php/www/init.php
9. Restart docker containers with docker-compose down && docker-compose up
10. Text BotFather in Telegram to update the user visible command list

Built With

• PHP Telegram Bot - A Telegram bot framework in PHP
• Telegram - Telegram Instant Messaging platform.
• IBM Watson - AI engine
• WeLive - Datasets sources
• Symfony - Web Framework

Versioning

We use SemVer for versioning. For the versions available, see the tags on this repository.

Authors

• Aitor Brazaola - Main developer - GitHub

License

This project is licensed under the GPL v3 License - see the LICENSE.md file for details

Acknowledgments

• docker-symfony - Szymon Skórczyński
• Example Bot Telegram - PHP Telegram Bot Developers
• Coffe & Daft Punk