Skip to content

Manual.md

Latest commit

 

History

History
238 lines (189 loc) · 6.07 KB

Manual.md

File metadata and controls

238 lines (189 loc) · 6.07 KB

Personal AI Assistant - Local Setup Guide (with Multilingual & Voice Recognition)

Note - This is not Actual Process is just a Manual for upcoming Modifaction and some guid for you To How to Create Your Own Assistant You Can try My or Use your Own After Creating.

1. Install Required Software & Libraries

a) Install Python & Dependencies

  • Ensure Python 3.8+ is installed: https://www.python.org/downloads/
  • Install required libraries: 
    pip install speechrecognition pyttsx3 openai torch transformers nltk spacy langdetect librosa numpy pyaudio opencv-python dlib face_recognition flask sqlite3

b) Setup Offline NLP & AI Models

  • Download spaCy's English model:

    python -m spacy download en_core_web_sm

  • Use OpenAI's GPT-2 locally (Hugging Face Transformers):

    from transformers import pipeline
chatbot = pipeline("text-generation", model="gpt2")
response = chatbot("Hello, how can I help you?", max_length=50)
print(response)

2. Setup Speech Recognition (Offline)

  • Install CMU PocketSphinx:

    pip install pocketsphinx

  • Code to recognize voice input:

    import speech_recognition as sr
recognizer = sr.Recognizer()
with sr.Microphone() as source:
    print("Listening...")
    audio = recognizer.listen(source)
    text = recognizer.recognize_sphinx(audio)
    print("You said:", text)

3. Setup Text-to-Speech (Offline)

  • Use pyttsx3 for offline voice response: 
    import pyttsx3
engine = pyttsx3.init()
engine.say("Hello Arya, how can I assist you?")
engine.runAndWait()

4. Multilingual Support (Offline)

  • Install langdetect for detecting the user's language:

    pip install langdetect

  • Code to detect spoken language and respond accordingly:

    from langdetect import detect

def detect_language(text):
    lang = detect(text)
    print(f"Detected Language: {lang}")
    return lang

user_input = "Hola, cómo estás?"
language = detect_language(user_input)

responses = {
    "es": "Estoy bien, ¿y tú?",
    "fr": "Je vais bien, et toi?",
    "en": "I'm fine, how about you?"
}

response = responses.get(language, "I'm fine, how about you?")
print(response)

5. Voice Recognition for Personalized Responses

  • Train the assistant to identify who is speaking and respond differently.

  • Install necessary libraries:

    pip install librosa numpy pyaudio

  • Code to recognize the speaker and respond accordingly:

    import speech_recognition as sr
import librosa
import numpy as np

# Load your voice sample
your_voice, sr_rate = librosa.load("your_voice.wav", sr=16000)
your_mfcc = np.mean(librosa.feature.mfcc(y=your_voice, sr=sr_rate, n_mfcc=40).T, axis=0)

# Function to identify speaker
def identify_speaker(audio_file):
    speaker_voice, sr_rate = librosa.load(audio_file, sr=16000)
    speaker_mfcc = np.mean(librosa.feature.mfcc(y=speaker_voice, sr=sr_rate, n_mfcc=40).T, axis=0)
    
    # Compare with your voice
    similarity = np.linalg.norm(your_mfcc - speaker_mfcc)
    
    if similarity < 50:  # Adjust this threshold as needed
        return "Arya"
    else:
        return "Guest"

# Recognize voice and respond accordingly
def recognize_and_respond():
    recognizer = sr.Recognizer()
    with sr.Microphone() as source:
        print("Listening...")
        audio = recognizer.listen(source)
        
        # Save audio to a temporary file
        with open("temp.wav", "wb") as f:
            f.write(audio.get_wav_data())
        
        speaker = identify_speaker("temp.wav")

        if speaker == "Arya":
            print("Hello Boss, how can I assist you?")
        else:
            print("Hello Guest, how may I help?")

# Run the assistant
recognize_and_respond()

6. Control ESP32 (Home Automation)

  • Install Flask for a local API:

    pip install flask

  • Create a simple API to send commands to ESP32:

    from flask import Flask
app = Flask(__name__)

@app.route('/turn_on_light')
def turn_on_light():
    # Send HTTP request to ESP32
    return "Light Turned ON!"

if __name__ == '__main__':
    app.run(port=5000)

  • ESP32 listens for API calls using MicroPython.

7. Face Recognition for Security

  • Install OpenCV & Dlib:

    pip install opencv-python dlib face_recognition

  • Code for face authentication:

    import face_recognition
import cv2

known_face = face_recognition.load_image_file("arya.jpg")
known_encoding = face_recognition.face_encodings(known_face)[0]

video = cv2.VideoCapture(0)
while True:
    ret, frame = video.read()
    face_locations = face_recognition.face_locations(frame)
    encodings = face_recognition.face_encodings(frame, face_locations)

    for encoding in encodings:
        match = face_recognition.compare_faces([known_encoding], encoding)
        if match[0]:
            print("Access Granted!")
        else:
            print("Access Denied!")

    cv2.imshow("Face Recognition", frame)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

video.release()
cv2.destroyAllWindows()

8. Local Database for Storing Tasks & Notes

  • Use SQLite for local storage:

    import sqlite3
conn = sqlite3.connect('assistant.db')
c = conn.cursor()
c.execute('''CREATE TABLE IF NOT EXISTS tasks (id INTEGER PRIMARY KEY, task TEXT)''')
conn.commit()
conn.close()

  • Add a task:

    def add_task(task):
    conn = sqlite3.connect('assistant.db')
    c = conn.cursor()
    c.execute("INSERT INTO tasks (task) VALUES (?)", (task,))
    conn.commit()
    conn.close()

9. Running Everything Locally

  • Launch Flask API for home automation: 
    python assistant_api.py
  • Run the AI assistant: 
    python assistant.py

Your personal AI assistant is now running completely offline on your local system with multilingual support and voice recognition! 🚀