A modular telegram bot written using GoLang. Utilizing microservices architecture
- GoLang
- Docker
- Python
- MongoDB
- gRPC
- AWS (Deployed)
This bot started as a small project to store class notes and keep assignment reminders on telegram. Slowly and gradually, a lot of features have been added to it
--- Bot (written on GoLang)
|
|
--- AI gRPC endpoint (written on Python)
|
|
--- MongoDB instance
|
|
Docker Daemon --- Mongo Express (to view the database)
|
|
--- ElasticSearch1 (load balanced)
|
|
--- ElasticSearch2 (load balanced)
The bot learns from the speaking habits of it's users and learns to talk back. It is surprisingly good, almost seems like a real person is talking back. The AI feature is handled by a python backend. The bot communicates with it using gRPC.
To use this feature, just start talking to the bot
/dadjoke - Sends a random Dad joke to the chat
/echo Hello World
replies:
Hello World
Echos back the message sent
/remind 25/02/2000 12:00 IST Rohan's Birthday
replies:
Rohan's Birthday
on 25/02/2000 12:00 AM
Keeps reminders
/save #<tag>
Saves a note with a tag. Also if a file is given, it is passed through a pipeline where OCR is run on it,
it is indexed and ready to be searched, like Google! but for your files.
/get #<tag>
Gets all notes with the given tag
/search #<tag> <query> - returns the file which contains the notes on query.
The query need not be exact, as the search engine uses fuzzy algorithms and replaces words with synonyms to find the perfect match.
/slap Rohan
replies:
Rohan hit Rohan around a bit with a trout
Slaps the person the command is used against
/stickerslap
Reply to a user with this command and the bot will slap them with a sticker
This bot can be deployed on any linux server with Docker installed. Using docker compose, it can also be deployed on services like AWS Elastic Beanstalk, DigitalOcean Application Service, etc
version: "3"
services:
bot:
image: rkrohk/gobot:latest
env_file: "./.env.gobot"
restart: always
mongo:
image: mongo
container_name: mongo
ports:
- 27017:27017
expose:
- 27017
volumes:
- mongodata:/data/db
chatbotapi:
image: rkrohk/chatbotapi
container_name: chatbotapi
hostname: chatbotapi
expose:
- 50051
environment:
- MONGO_URI=mongo
volumes:
- chatbotapi_data:/src/app/data
- nltk_data:/root/nltk_data
mongo-express:
image: mongo-express
environment:
- ME_CONFIG_MONGODB_SERVER=mongo
- ME_CONFIG_MONGODB_PORT=27017
ports:
- "8080:8081"
es01:
image: rkrohk/elasticsearch
container_name: es01
environment:
- node.name=es01
- cluster.name=es-docker-cluster
- discovery.seed_hosts=es02
- cluster.initial_master_nodes=es01,es02
- bootstrap.memory_lock=true
- "ES_JAVA_OPTS=-Xms512m -Xmx512m"
ulimits:
memlock:
soft: -1
hard: -1
volumes:
- data01:/usr/share/elasticsearch/data
ports:
- 9200:9200
es02:
image: rkrohk/elasticsearch
container_name: es02
environment:
- node.name=es02
- cluster.name=es-docker-cluster
- discovery.seed_hosts=es01
- cluster.initial_master_nodes=es01,es02
- bootstrap.memory_lock=true
- "ES_JAVA_OPTS=-Xms512m -Xmx512m"
ulimits:
memlock:
soft: -1
hard: -1
volumes:
- data02:/usr/share/elasticsearch/data
volumes:
mongodata:
nltk_data:
chatbotapi_data:
data01:
driver: local
data02:
driver: local
Along with the docker compose file, you also need a file named
.env.gobot to store secrets.
Structure of .env.gobot
BOT_TOKEN=<bot token>
BLOCKED_USER=<user ids of banned users>
OWNER=<your telegram user id, get it by using /id>