cjchirag7 / shopping-assistant

Shopping Assistant is a chatbot, which can assist consumers in deciding the right product, just like a virtual salesperson. This will bridge the gap between online and offline shopping.

Shopping Assistant

Index

• About
• Features
• Instructions for running locally
  • Initial steps
  • Method I (Using docker) [Recommended]
  • Method II (Without docker)
• Demo video
• Screenshots
• Tech Stack
• File Structure

About

Shopping Assistant is a chatbot, which can assist consumers in deciding the right product and bridge the gap between online and offline shopping.

Features

• Suggests products to the consumer depending upon his needs, just like a salesperson.
• Helps the consumer to virtually experience fashion products. E.g. If a consumer needs to try a T-shirt or a spectacle our shopping assistant gives him real time experience of how that product would look on him/her.
• Provides a summary of all the reviews about a product, which prevents users from doing the tedious job of going through hundreds of reviews of that product.

Instructions for running locally

Initial steps

1. Clone the repository by using the below command:

git clone https://github.com/cjchirag7/shopping-assistant

2. Download the 4 model files from here and copy them all to the directory shopping-assistant/services/api/. Please note that if you download the model_files folder in .zip format, then you need to extract all the files out of it and then copy them to the above directory.

Method I (Using docker) [Recommended]

Pre-requisites

1. Install Docker by looking up the docs
2. Install Docker Compose by looking up the docs

Note: If you are using Windows, make sure Docker Desktop is running.

Steps

1. Make sure you are in the root of the project (i.e., ./shopping-assistant/ folder).
2. Run docker-compose up to spin up the containers. If you are using Linux or Mac, you may need to use sudo for this command to work.
3. web-app would then be available locally at http://localhost:3000 , server at http://localhost:8000 and the API documentation would be available at http://localhost:8000/redoc

Method II (Without docker)

Running the Server

0. If you don't already have pipenv installed, install it using the following commands:

pip install --upgrade setuptools wheel
pip install --user pipenv

1. Activate the virtual environment in the api folder by using the following command:

cd services/api
pipenv shell

2. In the activated virtual environment, run the following command to install all the dependencies:

pipenv install

3. In the activated virtual environment, run the following command to run the API:

uvicorn main:app

4. The server would run at http://127.0.0.1:8000/ and the API documentation would be available at http://127.0.0.1:8000/docs

Running the Web App

1. Make sure you have node installed with version >= 14. Check using following command:

node -v

2. In the web-app folder, install all the dependencies using the following command:

cd services/web-app
npm install

3. In the web-app folder, run the React App using:

npm start

4. The web app would start running at http://localhost:3000

Demo video

• Available at https://www.youtube.com/watch?v=x_BFtcoaTks

Screenshots

Tech Stack

Server

• FastAPI
• Pytorch
• NLTK

Web App

• React.js
• Jeeliz

Others

• Docker

File structure

• Presentation has been attached as presentation.pdf
• Screenshots have been attached in screenshots/
• Training files have been attached as training-files/
• Server code has been attached in services/api/
• Frontend code has been attached in services/web-app/