PyGo Stock Portfolio Builder

Tool Design

Our program design is the combining structure of Golang and Python. There are three main files in our tool, which are portfolio.go, stockHandler.go, and web_interface.py. Additionally, out test cases are in portfolio_test.go. While portfolio.go and stockHandler.go are served as the back-end of the program, web_interface.py is the front-end. The motivation behind this project is to facilitate tracking and organizing an investment portfolio to equip the user with the most up to date information in order to make the best investment decisions. We were in need of tools like this since we started making investments and we thought it would be great to make one ourselves. The overall goal of this tool is to track a set of stock market positions for multiple portfolios, store the data, and display it using meaningful graphs and charts. In this tool, there are numbers of packages that are installed to support the program. For Golang, piquette/finance-go package is used to get access to the latest stock data. Additionally, we used some built-in Golang packages for encoding JSON files, implementing command line interface, and creating a unit test suite. For Python, we used pipenv and it's Pipfile to manage the installation of our dependencies and packages. Some of the used packages are pandas for data manipulation, shlex for formatting system commands before running, matplotlib.pyplot and altair for creating the visual plots such as bar charts and pie charts, and streamlit to implement graphical interface and a web application for our application. We also used Yfinance to validate stock tickers added by the user. The design of our tool involves a front end section that takes and validates user input. Following that, it runs specific system commands to send arguments to the backend portion. The back-end part uses object oriented programming to create and manipulate Go structures while also storing the information in json files. Overall, this organization allow us to combine the Python's elegant and user friendly graphical interface with Golang's efficient object oriented programming and data processing.

Installation

Running through docker

This repository contains a Dockerfile that could be built to run this tool or an already created image can be pulled from DockerHub.

• To build the image from Dockerfile, run the following in the cloned repository (including the . at the end):

docker build -t noorbuchi/pygo-portfolio-tracker .

OR

• Pull the image from Docker Hub using:

docker pull noorbuchi/pygo-portfolio-tracker

Once the image is done loading, you can use the following command to run the container

docker run -p 127.0.0.1:80:8501 -d -v /pathToJson/:/usr/src/app/portfolio noorbuchi/pygo-portfolio-tracker

Note: Make sure to replace pathToJson with your preferred path to store your portfolio json files. For example:

docker run -p 127.0.0.1:80:8501 -d -v /home/user/my-stock-portfolios/:/usr/src/app/portfolio noorbuchi/pygo-portfolio-tracker

The container will be running on localhost and the web application can be simply accessed by typing localhost in your favorite browser. All created json files will be saved to the directory you provided and you can mount that same directory again later to view the files contents.

Once you're done using the portfolio tracker, you can do the following to stop the docker container:

• run docker container ls to find the running container ID and name
• run docker stop container-id and replace container-id with the running container's ID

Local Installation Prerequisites

Python Environment

Python makes up all front-end aspects of this project. Specifically, web application and data retrieval are done using Python. To ensure that all Python components run smoothly, you will need to have:

• Python 3.8 or higher
• Pip, a Python package manager
• Pipenv, used to create a Python virtualenv and retrieve needed packages. This package can be installed using:

pip install pipenv

Golang Environment

Back-end operations and calculations, as well as data storage are made possible through Go. The following prerequisites are needed:

• A version of Go (1.14 or 1.15)
• finance-go package, can be installed using:

go get github.com/piquette/finance-go

Web Application

After all needed packages and tools are installed, the Python virtual environment needs to be created in the repository's root using:

pipenv install

This might take some time to locate all dependencies and create Pipfile.lock. However, once this is completed, you can run the tool's web application from the root of the repository using:

pipenv run streamlit run src/web_interface.py

Using Back-end Independently

stockHandler.go implements a command line interface that manages all operations and stores data. If you wish to use that without the web interface, you can follow the steps below:

• build all modules: go build src/stockHandler.go src/portfolio.go
• run the newly created stockHandler file using client line argument flags, example:
  • Deposit: ./stockHandler -f my_portfolio.json -o deposit -s 500
  • Withdraw: ./stockHandler -f my_portfolio.json -o withdraw -s 500
  • Buy: ./stockHandler -f my_portfolio.json -o buy -t TSLA -s 54.345 -p 20
  • Sell: ./stockHandler -f my_portfolio.json -o sell -t TSLA -s 54.345 -p 43

The test suite for the tool can be run using the following command or through the web application:

go test -v -cover ./src

Supported flags

• File Name

  • Flag: -f
  • Required: True
  • Type: string
  • Must contain the file format as .json

• Operation

  • Flag: -o
  • Required: False
  • Type: string
  • Possible arguments: buy, sell, withdraw, deposit

• Ticker

  • Flag: -t
  • Required: False
  • Type: string
  • Example argument: tsla, MSFT, amzn

• Shares

  • Flag: -s
  • Required: False
  • Type: float
  • In the case of using deposit or withdraw, the amount should be sent as a shares argument

• Price

  • Flag: -p
  • Required: False
  • Type: float
  • Used to input the price per share of the purchased/sold stock

• Refresh

  • Flag: -r
  • Required: False
  • Type: boolean
  • Could be sent alone to call the refresh function on the portfolio file

Program Implementation

There are three major files in our tool, each offers a distinct implementation and communicate with the others in specific ways. The first one is portfolio.go, which is the foundation of our program. The purpose of the file is to create the structure of the portfolio, stocks, and entry objects. It also contains functions for getting stock information, buying and selling stocks, deposit and withdrawing the cash, and most importantly, refreshing the data and creating a new portfolio. In these functions, buy and sell are the act of buying and selling stocks. It collect the stock ticker that users chooses, number of shares, and the price that these share were purchased or sold at. It then updated the Positions map in the portfolio object accordingly and recalculates totals. In the deposit and withdraw function, it tells users if they want to add or take the money from their portfolio. These functions will check the current amount of cash exiting in the portfolio and make the calculation to update the new amount. RefreshData is another important function that iterates through all values of the portfolio and calculates totals, percentages, and fetches latest stock prices. Following that, it updates the instance variables of the portfolio structure. This file is the foundation of the tool because it implements the objects structure and the basic functions needed to manipulate it.

The next file is stockHandler.go, which is also the back-end file. This file is served as the brain of the tool. stockHandler.go functions to handle and parse command line arguments as well, check for input error, handle file creation, and retrieve the data as a portfolio object. It also calls appropriate functions sending parameters based on user input. This section of the tool can be ran independently using the command line interface it implements.

web_interface.py is the front-end implementation that allows us to create the graphical interface of the program. The implementation uses Python and the Streamlit library to create and run the interface. Specifically, it allows users to check the updating status of their portfolio by showing the positions, cost, value, cash, gain/loss percentages, and total amount that they gain or loss with these stocks. All of this previously mentioned data is directly retrieved from the json files storing it. Additionally, the web interface sets up widgets such as buttons, text boxes, dropdown menus to take in user input. Once user input is added and the buttons to submit are clicked, web_interface.py determines the appropriate command to run and uses shlex and subprocess to do that. This causes a change in the data stored in the json file. The user can then refresh their page to display the latest information in a table show in the main page. In addition to sending commands to updated the data, the web interface provides informative visual representation of the data using altair and matplotlib. The interface allows the user to access the page with the graphs through a drop down menu. In that page, four charts can be seen, a pie chart showing the proportions that each stock makes up in the portfolio, a bar chart that compares the cost and gain/loss in dollars for each stock, a horizontal bar chart comparing the profitability percentage of each stock, and finally, a bar chart showing user activity since the portfolio was created. The last graph uses the history functionality that records and stores information about each operation done in the past. The history information can be seen in the History page of the web interface, where a table is shown. The web interface also contains functionalities for creating a new portfolio and switching between the portfolios stored in the Portfolio folder. The test suite can also be ran using the web interface and the results would be shown there. Overall, the web interface is able to handle all possible user input and display important messages that update the user and display any error messages.

Output Examples

Tool Screenshots

Test Suite Output

buchin@buchin:~/CS201/project-Noor-Anh$ go test -v -cover ./src
=== RUN   TestGetStockData
--- PASS: TestGetStockData (1.17s)
=== RUN   TestFileExists
=== RUN   TestFileExists/Test_web_interface.py
=== RUN   TestFileExists/Test_stockHandler.go
=== RUN   TestFileExists/Test_portfolio.go
=== RUN   TestFileExists/Test_../instructions.md
=== RUN   TestFileExists/Test_randomFile.json
--- PASS: TestFileExists (0.00s)
    --- PASS: TestFileExists/Test_web_interface.py (0.00s)
    --- PASS: TestFileExists/Test_stockHandler.go (0.00s)
    --- PASS: TestFileExists/Test_portfolio.go (0.00s)
    --- PASS: TestFileExists/Test_../instructions.md (0.00s)
    --- PASS: TestFileExists/Test_randomFile.json (0.00s)
=== RUN   TestPortDeposit
--- PASS: TestPortDeposit (0.00s)
=== RUN   TestPortWithdraw
--- PASS: TestPortWithdraw (0.00s)
=== RUN   TestPortBuy
--- PASS: TestPortBuy (0.08s)
=== RUN   TestPortDoubleBuy
--- PASS: TestPortDoubleBuy (0.21s)
=== RUN   TestPortSell
--- PASS: TestPortSell (0.21s)
PASS
coverage: 53.6% of statements
ok  _/home/buchin/CS201/project-Noor-Anh/src  (cached)  coverage: 53.6% of statements

The output shown above tests the backend operation done to the Portfolio structure and its functionalities. Some aspects could not be tested since they are dependent on changing stock market prices. The graphical interface was also tested with a variety of input to ensure that error messages are minimal.

Feedback

Thank you for checking out our tool! We would appreciate any feedback through the issue tracker and pull requests.

This tool was built by @noorbuchi and @tonyanh1710