A change in vision | Building Stocksimpy (Devlog 4)

#programming #python #opensource #learning

Imagine being able to test your strategy in less than 10 lines of Python code, rather than spending hours with other libraries. This is what will make `Stocksimpy` useful.

This is the 4th (technically 5th) post in my devlog series on building Stocksimpy, a lightweight Python backtesting library. Read the rest of the series here.

I’ve been developing Stocksimpy for about a month now. Along the way, I learned a lot, from how to properly document a library to how to write reusable code. Those were my original goals: I wanted to learn. But if I’m honest, learning for myself isn’t enough — I want to build something actually useful for others too.

This shift made me revisit the origin of Stocksimpy.

The idea came when I was building a simple stock prediction script. I could generate numbers and plots, but all I saw was just a bunch of data flowing and no real sense of whether the model was “working.” Was it actually successful, or just noise?

Naturally, I looked for existing Python backtesting tools. What I found was overwhelming: libraries that required setting up accounts, managing APIs, and losing myself among lines of configuration. These are powerful tools, but for quick experimentation, they felt like overkill. I didn’t want complexity — I wanted something lightweight, intuitive, and well-documented.

That’s how the idea Stocksimpy was born. A Python library designed to be simple enough for a weekend experiment, yet structured enough to scale as strategies get more advanced. From the start, I chose to build in public, updating the GitHub repo as the project grew.

But recently, I realised I was just re-implementing the same indicator logic that major libraries like TA-Lib already provide. That was never the purpose of Stocksimpy. So I am doing some reconfiguration on the skeleton of the library. Here is what it will look like:

indicators.py: Contains some simple common indicators for quick testing and development.
StockData: Holds the Pandas data frame related to stock data.
Portfolio: Contains the history of buy/sell, currently held stocks, and manages buy/sell operations.
Backtester: Where the user inputs their strategy and a StockData to test, where the main loop occurs.
Visualize: Visualizes the result and change of the total amount.

I finished coding a working draft for the first three, and will post a blog post about Portfolio soon.

As these are the main classes of my library, I am aiming for a workflow similar to the following:

def strategy():
  # This is where the user builds their own strategy
  skip

from stocksimpy import StockData, Visualize, Backtester # 1

# Import the data, supports SQLite, pd.dataframe, dict, yfinance, and more
data = StockData(df) # 2

# Create a Backtester object with $100,000 initial money
backtest = Backtester(data, strategy, initial_cap=100000) # 3

# Runs it
backtest.run_backtest() # 4

# Visualizes the cap change
Visualize.visualize(backtest) # 5

So if this works as I expected, with only 5 lines of code it is possible to run a backtest for a strategy.

If there is anything you want me to add, or any suggestions, please reach out to me on my socials, or comment it below 👇. It really helps with improving the library.

If you would like to support me: