holohup / div_bot_2.0

microservices bot version

ZioNet Homework Microservices Application

TL;DR

The application retrieves current stock and futures prices and calculates future dividends based on the futures expiration date and the current discount rate.

It uses a microservices architecture and provides an API exposed to the outside world. There is also a sample Telegram bot to demonstrate the service's functionality.

The Bot

For demonstration purposes, I have created a small Telegram bot that acts as a frontend for the backend in production. Feel free to add @zionnet_hw_bot to your Telegram to see that the app is working. It does not store any user details. The bot accepts two types of commands:

• /list - lists tickers with at least one available future
• TICKER - send any exchange ticker from the previous list in any combination of upper/lowercase, like 'sber', 'TATNP,' or any other. The bot will poll the backend and format the JSON to something resembling a table in Telegram. The bot is not part of the docker-compose.yml since it can be run anywhere. Its source code is included in the project. Currently, the service and the bot are running 24/7 on a Raspberry Pi 4 4GB.

Installation

• Clone the repository:

git clone https://github.com/holohup/div_bot_2.0.git && cd div_bot_2.0

• Launch Docker Compose:

docker compose up [-d]

After some rustling, the app with all its microservices should be running.

Another External API

The application uses an external API - Tinkoff Broker. If you happen to have its API key, you can add it to the TCSApiAccessor/.env.example, and the application will work correctly. Without it, it provides data from fixtures - everything, including tests, runs smoothly, but you cannot trust the data it provides. The Telegram bot is connected to a backend that uses the real key, so the bot responses are legitimate.

How to Test

API

The two endpoints are available via a Swagger interface at http://127.0.0.1:8005/docs/.

Integration Tests

!NB Make sure you have Docker Compose running.

In the app root folder, create a virtual environment on a machine with Python installed, activate it, install the dependencies, and launch pytest - all of them should pass:

python3 -m venv venv && source venv/bin/activate && pip install -r requirements.txt && pytest

Structure

The structure above is self-explanatory. I'll write more details after I get some feedback from real users who wish to know more about the structure. However, there's one important point:

log.txt

Since the logging server relies on Queue, it works asynchronously and adds the results of dividend estimation to the log.txt file in the root folder of the project. It's mapped from the LogAccessor, so you should see the changes in the log.

To-Do

• Decouple business logic from management - create a financial engine underneath the manager
• Increase test coverage, especially for data handling and None values
• If the project is to be developed further, add more docstrings and type annotations
• Add health check endpoints for Docker Compose
• Try to use MarkdownV2 and with mono-width font in a telegram bot for a neater table
• Implement CI/CD using Github Actions: flake8, isort, unit and integration tests, images creation and push to Dockerhub, optionally SSH for deployment.
• DAPRize it.