Stockbot

This repository contains the code of a Telegram bot that gets information about stocks.

Supported commands

The bot supports the following commands:

1. /start - show the description of what the bot can do
2. /help - show the list of commands the bot supports
3. /price {symbol} - get price stats
4. /return {symbol} - get return stats
5. /vol {symbol} - get volatility stats
6. /all {symbol} - get price, return, and volatility stats

For a detailed description of the response of each command, please see: commands.md

Monitor mode

In addition to the commands, the bot also has a monitor mode which is responsible for two things:

1. Send an alert when there is a price anomaly on a symbol of our portfolio.
2. Send a daily report about our portfolio performance.

For a detailed description of what the monitor mode does, please see: monitor.md

High level design

This is the high level design of Stockbot:

Install

1. python -m venv .venv
2. source .venv/bin/activate
3. make install
4. make build

Test

To execute the unit tests please run the following command: make test

Coverage

To see the code coverage report please execute the following commands:

1. make coverage
2. open htmlcov/index.html

Run

First define the required environment variables:

1. export TELEGRAM_BOT_TOKEN='<bot-token-here>'
2. export CHANNEL_ID='<channel-id-here>'
3. export SYMBOLS='<symbol1>,<symbol2>,...,<symbolN>'

Please enter symbols that are valid in Yahoo Finance.

Run locally

We can run the bot in two modes locally:

1. Polling mode: make poll
2. Monitor mode: make monitor

Maintenance

The project has three directories at the root level:

1. docs: contains files related to documentation.
2. infra: contains CDK files that define the infrastructure.
3. src: contains the source code and tests of the bot.

Source code

The code is organized in the following way:

1. We have components, each component has a single responsibility.
2. Each component has its own folder.
3. The folder of each component has its source code and tests.

We have the following components:

1. analyst: it is responsible for performing the analysis over the data.
2. bot: it is responsible for receive commands and return answers.
3. common: it is responsible to hold common types and data structures.
4. download: it is responsible for downloading the data we need to analyst.

In addition to the components we have 4 files in the root of the project:

1. commands: contains the commands supported by the bot.
2. monitor: lambda to monitor a list of symbols.
3. poll: code to get new messages via polling (polling Telegram).
4. push: lambda to get new messages via push (webhooks sent by Telegram).

Project structure

To visualize the project structure, please execute the following command:

tree -I 'venv|__pycache__|test_files|htmlcov|cdk.out'

Deploy

We use AWS CDK (infrastructure as code) + AWS API Gateway + AWS Lambda to deploy the bot. The details are documented here.

Run the following command to create the infrastructure and deploy the code:

make deploy

If we want to delete the infrastructure created, please run this command:

make destroy