GitLocker: The Coding Marketplace

Description:

PerceptronRabiaWadhwa 0.0.2

Perceptron
References

Python Package Creation Docs
Github Actions - Publishing to package registries

CI/CD.yml code
# name of the workflow
name: Upload Python Package

#when to execute - when you push changes on "main" branch
on:
push:
branches:
- main
# - dev => add for more branches if needed

# what to execute - lists jobs
jobs:
deploy: # first job ,you can list more jobs as well
runs-on: ubuntu-latest # CREATE UBUNTU ENVIRONMENT
steps:
- uses: actions/checkout@v2
- name: Set up Python #sets up python
uses: actions/setup-python@v2
with:
python-version: '3.7' # python version needed
- name: Install dependencies
run: |
python -m pip install --upgrade pip
pip install build
- name: Build package
run: python -m build
- name: Publish package
uses: pypa/gh-action-pypi-publish@27b31702a0e7fc50959f5ad993c78deac1bdfc29
with:
user: __token__
password: ${{ secrets.PYPI_API_TOKEN }}

How to use package
from perceptron_package.perceptron_class import Perceptron
import pandas as pd

def prepare_data(df):
""" Used to separate dependent and independent features

Args:
df (pd.dataframe): pandas dataframe

Returns:
tuple: returns tuple of dependent & independent variables
"""
X = df.drop("y",axis=1)
y = df["y"]
return X,y

def main(data ,eta,epochs):
df = pd.DataFrame(data)
df # Shape = (4,3)

X,y = prepare_data(df)

model = Perceptron(eta=eta, epochs=epochs) # Creating object of class Perceptron
model.fit(X, y) # Weights in last epoch are considered as final weights for prediction

_ = model.total_loss() # last Epoch's Sum of Errors , '_' indicates dummy variable

if __name__ == '__main__': # define entry point of program execution
AND = {"x1":[0,0,1,1],
"x2":[0,1,0,1],
"y" :[0,0,0,1]
}

ETA = 0.3 # between 0 and 1
EPOCHS = 10
main(AND,ETA,EPOCHS)