Breast-Cancer-Prediction-Docs.txt

Sagemaker links - 

https://towardsdatascience.com/a-practical-guide-to-mlops-in-aws-sagemaker-part-i-1d28003f565

https://aws.amazon.com/getting-started/hands-on/build-train-deploy-machine-learning-model-sagemaker/



---------------------------------------------------------
Build ML model using sagemaker Studio notebook - 

Yiu tube - https://www.youtube.com/watch?v=1iSiN4sVMjE&t=1900s
https://github.com/seanpmorgan/studio-notebook-webinar 
https://github.com/aws-samples/sagemaker-studio-lifecycle-config-examples/blob/main/scripts/install-lsp-features/on-jupyter-server-start.sh

pip install jupyterlab-code-formatter black
%pip install --ignore-installed --upgrade tensorflow
%pip install --upgrade boto3 sagemaker -U
pip install scikit-learn --upgrade




Go to lambda on AWS
name to lambda function
runtime - python 3.6
Architecture - x86_64

Create fucntion 
copy the code form AWS
deploy
go to
-----------------------------------------------------------------------------------------------------------------------

Breast Cancer Prediction

# End to end sagemaker with lambda API

https://sagemaker-examples.readthedocs.io/en/latest/introduction_to_applying_machine_learning/breast_cancer_prediction/Breast%20Cancer%20Prediction.html
https://aws.amazon.com/blogs/machine-learning/call-an-amazon-sagemaker-model-endpoint-using-amazon-api-gateway-and-aws-lambda/
You Tube - https://www.youtube.com/watch?v=stD47vPDadI

Creatr Sagemaker Instance
instance type - ml.t.medium
Permission and encryption - IAM role - Create new role - Any S3 bucket - Create notebook instance
click on Open Jupyter 
Go to Sagemaker Examples
use Breast Cancer Prediction.ipynb - Create Copy

It will open the Breast Cancer Prediction notebook
%pip install --upgrade boto3 sagemaker -U
Run all the cells.
Once all cells have been run and endpoint created go to Sagemaker Lambda
Create Function- Give name - 
runtime - Python 3.9
replace the below code in lambda fucntion:

import os
import io
import boto3
import json
import csv

# grab environment variables
ENDPOINT_NAME = os.environ['ENDPOINT_NAME']
runtime= boto3.client('runtime.sagemaker')

def lambda_handler(event, context):
    print("Received event: " + json.dumps(event, indent=2))
    
    data = json.loads(json.dumps(event))
    payload = data['data']
    print(payload)
    
    response = runtime.invoke_endpoint(EndpointName=ENDPOINT_NAME,
                                       ContentType='text/csv',
                                       Body=payload)
    print(response)
    result = json.loads(response['Body'].read().decode())
    print(result)
    pred = int(result['predictions'][0]['score'])
    predicted_label = 'M' if pred == 1 else 'B'
    
    return predicted_label

Deploy
go to configuration
Go to Variable Environment - edit - Create new environment
Key - ENDPOINT_NAME
Value - Genter the name of enpoint (Go to sagemaker inference to copy the endpoint name)

Go to IAM - role - select the lambda funtcion role-permission policies - select policy AWSLambdaBasicExecutionRole-b28097

edit policy- go to JSON and paste below code

       {
            "Sid": "VisualEditor0",
            "Effect": "Allow",
            "Action": "sagemaker:InvokeEndpoint",
            "Resource": "*"
        }

review-save changes


Go to API gateway in AWS Console
REST API - Build - ok - 
Choose the protocol - REST
Create New API - New API
Settings - API name - give name
	Description - give description
	create API

Go to Actions - Create Resource -give resource name- Create Resource
Go to Actions - Create Method -select POST - Tick the checkbox next to it
Integration type - Lambda fuction
Lambda Function - select the name of your lambda fucntion created
Save -Ok
Test
in Request Body paste below data:
{"data": "1257815, 5, 1, 3, 1, 2, 2, 1, 1, 6, 4, 3, 2, 5, 5, 6, 5, 4, 4, 23, 6, 3, 2, 7, 8, 8, 3, 2, 1, 6"} 

Test
It will give you a prediction for the given input as "B"


Go to Action-  Deploy API
Deployment stage - [New Stage]
Stage Name - PROD
Stage Description - Production Environment
Deployment Sescription - 
Deploy 

It will create Invoke URL

Go to AWS console- Create T2.micro instance - 

sudo su
sudo apt update
sudo apt upgrade
apt install python3-pip
sudo pip install --upgrade boto3
python3
import json
import requests
r=requests.post("https://ntvhjus37g.execute-api.us-east-1.amazonaws.com/PROD/api-ml-model", data=json.dumps({"data": "1257815, 5, 1, 3, 1, 2, 2, 1, 1, 6, 4, 3, 2, 5, 5, 6, 5, 4, 4, 23, 6, 3, 2, 7, 8, 8, 3, 2, 1, 6"})) 
r.content

It will give you a prediction as "B"