pip install -r requirements.txt

Importing Google Contacts

To enable sending messages and making calls to your contacts, you need to import your Google contacts in CSV format and run the following commands in db.py (SQLite query):

import csv
import sqlite3

# Connect to SQLite database
con = sqlite3.connect('jarvis.db')
cursor = con.cursor()

# Specify the column indices you want to import (0-based index)
# Example: Importing the 1st and 34th columns (1st column contains the name and 34th contains the contact number of the person)
desired_columns_indices = [0, 33]

# Read data from CSV and insert into SQLite table for the desired columns
with open('contacts.csv', 'r', encoding='utf-8') as csvfile:
    csvreader = csv.reader(csvfile)
    for row in csvreader:
        selected_data = [row[i] for i in desired_columns_indices]
        cursor.execute('''INSERT INTO contacts (id, name, mobile_no) VALUES (null, ?, ?);''', tuple(selected_data))

# Commit changes and close connection
con.commit()
con.close()

Authenticating with Hugging Face

For authentication to Hugging Face via the cookies.json file, follow these steps:

1. Go to Hugging Face Chat.
2. Use a browser extension like Cookie-Editor to create, edit, and delete cookies for the current tab.
3. Click on the extension after navigating to the Hugging Face Chat website and copy the cookies in JSON format.
4. Save the cookies in a cookies.json file as shown in the folder structure below.

Folder Structure

Here's the folder structure for the JARVIS AI Assistant project:

JARVIS-AI-ASSISTANT/
├── engine/
│   ├── __pycache__/
│   ├── command.py
│   ├── config.py
│   ├── cookies.json (from cookie-editor)
│   ├── db.py
│   ├── features.py
│   ├── helper.py
│   ├── tempCodeRunnerFile.py
│   ├── test.py
├── frontend/
│   ├── assets/
│   ├── controller.js
│   ├── index.html
│   ├── main.js
│   ├── script.js
│   ├── style.css
├── jarvisenv/
├── .gitignore
├── contacts.csv (from google contacts)
├── jarvis.db
├── main.py
├── requirements.txt
├── run.py

Architecture

• Voice Recognition: Powered by advanced NLP models from Hugging Face.
• Text-to-Speech: Converts AI responses into natural-sounding speech.
• Persistent Storage: Uses SQLite for storing chat history and commands.
• Hotword Detection: Always-on listening for the hotword "JARVIS".

Example Commands

• "JARVIS, what's the weather today in Mumbai?" (usses llama3 model integrated using huggingface moudle hugchat)
• "Send message to Srijan: I'll be late today." (Automatically sends message to that person on whatsapp)
• "Voice call Srijan." (Searches for the username Srijan in the contacts file and starts video calling the person)
• "Video call Srijan."
• "JARVIS, play the latest news on YouTube." (spins up a youtube video on default web browser of the system)
• "JARVIS, play Counting stars by OneRepublic on YouTube."

Contributing

Feel free to fork this repository and contribute by submitting a pull request. For major changes, please open an issue to discuss what you would like to change.

Preview:

Type or speak the prompt:

The it would automatically start the browser and play video on youtube.

Chatgpt like interface, using hugchat module to integrate huggingface, model used as of now is Llama3, the user has the ability to switch between the models by configuring settings interface or simply asking JARVIS to do it.

Prompt: Explain string theory like I am 10