Topsis-Niyati-102303356 1.0.1

A Python package for TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) multi-criteria decision analysis

TOPSIS Multi-Criteria Decision Analysis System

End-to-end implementation of the TOPSIS algorithm — from a pip-installable Python package to a deployed REST API with a web interface.

---

What This Project Covers

Part	What I Built	Tech
I	TOPSIS algorithm from scratch + CLI tool	Python, NumPy, Pandas
II	Published Python package to PyPI	setuptools, pip, PyPI
III	REST API + Web Interface	Flask, HTML/CSS/JS

---

The Algorithm — TOPSIS

TOPSIS ranks decision alternatives by measuring how close each is to the ideal best and how far from the ideal worst.

Steps:

1. Normalize the decision matrix
2. Apply criterion weights
3. Compute positive and negative ideal solutions
4. Calculate Euclidean distances to each ideal
5. Compute performance scores → rank

Score(i) = D_worst(i) / (D_best(i) + D_worst(i))

---

Part I — CLI Tool

Install

pip install Topsis-Niyati-102303356

Usage

topsis <InputFile.csv> <Weights> <Impacts> <OutputFile.csv>

# Example
topsis data.csv "1,1,1,2" "+,+,-,+" result.csv

Input CSV format

Model,Storage,Price,Camera,Battery
M1,16,2,3,4
M2,16,2.5,4,4
M3,32,3,4,5

• Column 1: alternative names
• Columns 2+: numeric criteria values

Output CSV format

Original data + two new columns: Topsis Score and Rank

Validations

• Correct number of CLI arguments
• File exists and is readable
• Input file has ≥ 3 columns
• Columns 2–n are numeric only
• Weights and impacts count matches column count
• Impacts are + or - only
• Weights are positive numbers

---

Part II — PyPI Package

Package: Topsis-Niyati-102303356

Use as a library

import pandas as pd
from topsis import topsis

data    = pd.read_csv("data.csv")
weights = [1, 1, 1, 2]
impacts = ['+', '+', '-', '+']

result = topsis(data, weights, impacts)
print(result[['Model', 'Topsis Score', 'Rank']])

---

Part III — REST API + Web Interface

Run backend locally

pip install flask flask-cors pandas numpy
python app.py
# → http://localhost:5000

Open web interface

Just open index.html in any browser — no build step needed.

API — POST /topsis

curl -X POST http://localhost:5000/topsis \
  -F "file=@data.csv" \
  -F "weights=1,1,1,2" \
  -F "impacts=+,+,-,+" \
  -F "email=[e-mail]" \
  --output result.csv

---

Project Structure

topsis-project/
├── topsis/
│   ├── __init__.py       # Package init
│   ├── core.py           # TOPSIS algorithm (NumPy/Pandas)
│   └── cli.py            # Command-line entry point
├── app.py                # Flask REST API
├── index.html            # Web interface
├── data.csv              # Sample input
├── setup.py              # PyPI config
├── requirements.txt
└── README.md

---

Skills Demonstrated

• Algorithm implementation from academic paper → production code
• Python package development and PyPI publishing
• REST API design with Flask
• Vectorized computation with NumPy
• End-to-end delivery: CLI → package → API → UI

---

Dependencies

pandas>=1.3.0
numpy>=1.21.0
flask>=2.0.0
flask-cors>=3.0.0

License

MIT