Bangkit Capstone Project

This is repository aplication of CoffeeScape

Please change the branch to cc, ml, or android to see each path repository!

Our Design app

Our Member

Member	Student ID	Path
Irfan Fadli Nugraha	M008BSY0620	Machine Learning
Hana Dewi Shoviyah	M283BSX0521	Machine Learning
Enas Erliana Zakiya Yudhana	M008BSX0125	Machine Learning
Muhamad Fihris Aldama	C296BSY4031	Cloud Computing
Rayya Ruwa’im Nafie	C296BSY3695	Cloud Computing
Talitha Bertha Arvyandita	A296BSX2694	Android Development
Lutfi Nur Rohmah	A015BSX2004	Android Development

Natural Language Processing, Collaborative Filtering, and Mood Based Recommendation.

This repository contains three scripts for different tasks: sentiment analysis using deep learning, collaborative filtering using cosine similarity, and mood based recommendation using switch case to initiate randomly generated value based on each mood group.

Sentiment Analysis with Deep Learning.

Requirement

Python 3.x
TensorFlow 2.x
Pandas
NumPy
Scikit-Learn

Instruction

1. Install Dependencies

pip install tensorflow pandas numpy scikit-learn

2. Download The Dataset

For the sentiment analysis task, download the 'train.csv' file and place it in the same directory as the script.

3. Run the Sentiment Analysis Script

NLP.py This script loads the dataset, preprocesses the text data, creates a deep learning model for sentiment analysis, and saves the trained model as 'NLP_model.h5'.

4. Testing the trained model.

Edit the test_text variable in the script with your own text and run the script to get sentiment predictions.

Model Details

The deep learning model architecture consists of an embedding layer, bidirectional LSTM layer, and several dense layers.
The model is trained using binary crossentropy loss and the Adam optimizer.

Collaborative Filtering with Cosine Similarity.

Requirement

Python 3.x
Pandas
NumPy
Scikit-Learn

Instruction

1. Install Dependencies

pip install pandas numpy scikit-learn

2. Download The Collaborative Filtering Dataset

For the collaborative filtering task, download the 'NEW_Dataset Capstone.xlsx - Rating Data (5).csv' file and place it in the same directory as the script.

3. Run The Collaborative Filtering Script

python collaborative_filtering.py This script reads the dataset, computes item similarity using cosine similarity, generates new user ratings, and recommends top items for the new user.

4. Save the model

The script saves the item similarity matrix using pickle as 'finalized_model.pkl'.

Model Details

Collaborative filtering is implemented using cosine similarity between items based on user ratings.
The script demonstrates how to recommend items for a new user.

Mood Based Recommendation

Requirements

JavaScript-enabled environment (browser, Node.js, etc.)

Instruction

1. Get the data

Get the data from the dataset for mood based and put them inside the list of each variables. Ensure you have access to the dataset containing mood-based drink information.

2. Run the script

Open the HTML file containing the script in a browser or execute the script using Node.js.

3. Enter Your Mood

When prompted, enter your mood (happy, sad, lonely, or bored).

4. View the Suggested Drink

The script will output a suggestion based on your mood.

Deploy ML Model

Server Requirements

Python - version 3.8 or above.

Installation

Clone This Repo

git clone -b ml https://github.com/fihrisaldama015/Capstone_CoffeScape.git
cd Capstone_CoffeScape/API_FLASK

Clone the ml branch & go to the API_FLASK folder directory

Install Dependencies

pip install --user tensorflow
pip install --user flask
pip install --user pandas
pip install --user pickle
pip install --user numpy

Run the ML API

$ python3 coffeescape.py

the server run on port 5000

CoffeeScape API

Documentation

API DOCUMENTATION

Server Requirements

Node.js - version 18.18.0 or above.

Link Download Node.js => Click This Link to Download

Installation

Clone This Repo

git clone -b cc https://github.com/fihrisaldama015/Capstone_CoffeScape.git
cd Capstone_CoffeScape

Install Dependencies

npm install

wait this installation proccess to complete. it takes 3-5 minutes.

Create ENV file

copy .env.example and rename the file to .env

JWT_SECRET_KEY= #secret
DATABASE_URL= #https://$PROJECT_ID.firebaseio.com
ML_API_ENDPOINT= #https://$PROJECT_ID.$REGION.appspot.com or http://$EXTERNAL_IP:$PORT
APP_URL= #https://$PROJECT_ID.web.app or http://$EXTERNAL_IP:$PORT

edit the file and use your own key and url

Generate Service Account

Create your Firestore database at Firestore Page, select project if you haven't.
Go to Service Account Page, then select your project (ex: capstone) if you haven't choose project yet.
Select one of the service account that have firebase-adminsdk in the beginning of the service account name.
Move to tab KEYS, click ADD KEY then select Create New Key
for key type select JSON, then click CREATE, the JSON file is downloaded to your local computer
Go to the downloaded JSON file directory, copy or move the file to the previous Capstone_CoffeeScape folder

Run the API

$ npm run dev

you should see like this when the server run successfully

> capstone@1.0.0 dev
> nodemon src/server.js

[nodemon] 3.0.1
[nodemon] to restart at any time, enter `rs`
[nodemon] watching path(s): *.*
[nodemon] watching extensions: js,mjs,cjs,json
[nodemon] starting `node src/server.js`
Server running on port 9000