Create app.py

app.py

@@ -0,0 +1,84 @@
+import gradio as gr
+import numpy as np
+import cv2
+import librosa
+import moviepy.editor as mp
+import speech_recognition as sr
+import tempfile
+import wave
+import os
+import tensorflow as tf
+from tensorflow.keras.preprocessing.text import tokenizer_from_json
+from tensorflow.keras.models import load_model, model_from_json
+from sklearn.preprocessing import StandardScaler
+from tensorflow.keras.preprocessing.sequence import pad_sequences
+import nltk
+from nltk.corpus import stopwords
+from nltk.stem import WordNetLemmatizer
+import pickle
+import json
+from transformers import pipeline
+
+nltk.download('punkt')       # Tokenizer
+nltk.download('wordnet')     # WordNet lemmatizer
+nltk.download('stopwords')   # Stopwords
+
+# Load text emotion model
+with open('model_architecture_for_text_emotion_updated_json.json', 'r') as json_file:
+    model_json = json_file.read()
+text_model = model_from_json(model_json)
+text_model.load_weights("model_for_text_emotion_updated(1).keras")
+
+# Load tokenizer
+with open('tokenizer.json') as json_file:
+    tokenizer_json = json.load(json_file)
+tokenizer = tokenizer_from_json(tokenizer_json)
+
+# Load LLM chatbot (replace with LLama or another LLM of your choice)
+chatbot = pipeline("text-generation", model="facebook/blenderbot-3B")
+
+# Initialize NLTK
+lemmatizer = WordNetLemmatizer()
+stop_words = set(stopwords.words('english'))
+
+def preprocess_text(text):
+    tokens = nltk.word_tokenize(text.lower())
+    tokens = [word for word in tokens if word.isalnum() and word not in stop_words]
+    lemmatized_tokens = [lemmatizer.lemmatize(word) for word in tokens]
+    return ' '.join(lemmatized_tokens)
+
+# Transcribe audio and get emotion
+
+def transcribe_audio(audio_file):
+    recognizer = sr.Recognizer()
+    with sr.AudioFile(audio_file) as source:
+        audio_record = recognizer.record(source)
+        text = recognizer.recognize_google(audio_record)
+    
+    pre_text = preprocess_text(text)
+    title_seq = tokenizer.texts_to_sequences([pre_text])
+    padded_title_seq = pad_sequences(title_seq, maxlen=35, padding='post', truncating='post')
+    inp1 = np.array(padded_title_seq)
+    text_prediction = text_model.predict(inp1)
+    
+    mapping = {0: "anger", 1: "disgust", 2: "fear", 3: "joy", 4: "neutral", 5: "sadness", 6: "surprise"}
+    max_index = text_prediction.argmax()
+    return text, mapping[max_index]
+
+# Chatbot response
+
+def chatbot_response(audio_file):
+    user_input, emotion = transcribe_audio(audio_file)
+    response = chatbot(user_input, max_length=100, num_return_sequences=1)[0]['generated_text']
+    return f"Detected Emotion: {emotion}\nChatbot: {response}"
+
+# Create Gradio Interface
+iface = gr.Interface(
+    fn=chatbot_response,
+    inputs=gr.Audio(source="microphone", type="filepath"),
+    outputs="text",
+    title="Emotion-Aware AI Chatbot",
+    description="Speak into the microphone, and the chatbot will analyze your emotion and respond accordingly."
+)
+
+iface.launch()