opinder2906/Software_engineering

1import torch2import torch.nn as nn3import torch.nn.functional as F4import random5from textblob import TextBlob6import pandas as pd7import requests8from io import StringIO9import gradio as gr10import speech_recognition as sr11import json12 13# ----- Dummy Model and vocab -----14vocab = {'<PAD>': 0, '<UNK>': 1, 'i': 2, 'am': 3, 'feeling': 4, 'sad': 5, 'happy': 6, 'angry': 7, 'love': 8, 'stressed': 9, 'anxious': 10}15MAX_LEN = 1616 17class DummyLabelEncoder:18    def __init__(self):19        self.classes_ = ['sadness', 'anger', 'love', 'happiness', 'neutral']20    def transform(self, x): return [self.classes_.index(i) for i in x]21    def inverse_transform(self, x): return [self.classes_[i] for i in x]22 23le = DummyLabelEncoder()24 25class DummyModel(nn.Module):26    def __init__(self):27        super().__init__()28        self.embedding = nn.Embedding(len(vocab), 8)29        self.fc = nn.Linear(8, len(le.classes_))30    def forward(self, x):31        x = self.embedding(x)32        x = x.mean(dim=1)33        return self.fc(x)34 35model = DummyModel()36 37def preprocess_input(text):38    tokens = text.lower().split()39    encoded = [vocab.get(token, vocab['<UNK>']) for token in tokens]40    padded = encoded[:MAX_LEN] + [vocab['<PAD>']] * max(0, MAX_LEN - len(encoded))41    return torch.tensor([padded], dtype=torch.long).to(next(model.parameters()).device)42 43# ----- Load CSV from Google Drive -----44file_id = "1yVJh_NVL4Y4YqEXGym47UCK5ZNZgVZYv"45url = f"https://drive.google.com/uc?export=download&id={file_id}"46response = requests.get(url)47csv_text = response.text48 49if csv_text.strip().startswith('<'):50    raise Exception("ERROR: Google Drive link is not returning CSV! Check your sharing settings.")51 52solutions_df = pd.read_csv(StringIO(csv_text), header=0, on_bad_lines='skip')53 54used_solutions = {emotion: set() for emotion in solutions_df['emotion'].unique()}55 56# ----- Data and responses -----57negative_words = [58    "not", "bad", "sad", "anxious", "anxiety", "depressed", "upset", "shit", "stress",59    "worried", "unwell", "struggling", "low", "down", "terrible", "awful",60    "nervous", "panic", "afraid", "scared", "tense", "overwhelmed", "fear", "uneasy"61]62 63responses = {64    "sadness": [65        "It’s okay to feel down sometimes. I’m here to support you.",66        "I'm really sorry you're going through this. Want to talk more about it?",67        "You're not alone — I’m here for you."68    ],69    "anger": [70        "That must have been frustrating. Want to vent about it?",71        "It's okay to feel this way. I'm listening.",72        "Would it help to talk through it?"73    ],74    "love": [75        "That’s beautiful to hear! What made you feel that way?",76        "It’s amazing to experience moments like that.",77        "Sounds like something truly meaningful."78    ],79    "happiness": [80        "That's awesome! What’s bringing you joy today?",81        "I love hearing good news. 😊",82        "Yay! Want to share more about it?"83    ],84    "neutral": [85        "Got it. I’m here if you want to dive deeper.",86        "Thanks for sharing that. Tell me more if you’d like.",87        "I’m listening. How else can I support you?"88    ]89}90 91# --- Helper functions ---92 93def correct_spelling(text):94    return str(TextBlob(text).correct())95 96def get_sentiment(text):97    return TextBlob(text).sentiment.polarity98 99def is_negative_input(text):100    text_lower = text.lower()101    return any(word in text_lower for word in negative_words)102 103def get_unique_solution(emotion):104    available = solutions_df[solutions_df['emotion'] == emotion]105    unused = available[~available['solution'].isin(used_solutions[emotion])]106    if unused.empty:107        used_solutions[emotion] = set()108        unused = available109    solution_row = unused.sample(1).iloc[0]110    used_solutions[emotion].add(solution_row['solution'])111    return solution_row['solution']112 113def get_emotion(user_input):114    if is_negative_input(user_input):115        return "sadness"116    sentiment = get_sentiment(user_input)117    x = preprocess_input(user_input)118    model.train()119    with torch.no_grad():120        probs = torch.stack([F.softmax(model(x), dim=1) for _ in range(5)])121        avg_probs = probs.mean(dim=0)122        prob, idx = torch.max(avg_probs, dim=1)123    pred_emotion = le.classes_[idx.item()]124    if prob.item() < 0.6:125        return "neutral"126    if sentiment < -0.25 and pred_emotion == "happiness":127        return "sadness"128    if sentiment > 0.25 and pred_emotion == "sadness":129        return "happiness"130    return pred_emotion131 132def audio_to_text(audio_file):133    if audio_file is None:134        return ""135    recog = sr.Recognizer()136    with sr.AudioFile(audio_file) as source:137        audio = recog.record(source)138    try:139        text = recog.recognize_google(audio)140        return text141    except Exception:142        return ""143 144# ----- Chat function -----145GLOBAL_CONVO_HISTORY = []146USER_FEEDBACK_STATE = {}147 148def emoti_chat(audio, text, history_json=""):149    if text and text.strip():150        user_input = text151    elif audio is not None:152        user_input = audio_to_text(audio)153    else:154        user_input = ""155    if not user_input.strip():156        return "Please say something or type your message.", json.dumps(GLOBAL_CONVO_HISTORY[-5:], indent=2), ""157 158    user_input = correct_spelling(user_input)159 160    exit_phrases = ["exit", "quit", "goodbye", "bye", "close"]161    if user_input.lower().strip() in exit_phrases:162        return "Take care! I’m here whenever you want to talk. 👋", json.dumps(GLOBAL_CONVO_HISTORY[-5:], indent=2), gr.update(visible=False)163 164    user_id = "default_user"165    state = USER_FEEDBACK_STATE.get(user_id, {"emotion": None, "pending": False})166 167    if state["pending"]:168        feedback = user_input.lower().strip()169        GLOBAL_CONVO_HISTORY[-1]["feedback"] = feedback170        if feedback == "no":171            suggestion = get_unique_solution(state["emotion"])172            reply = f"Here's another suggestion for you: {suggestion}\nDid this help? (yes/no/skip)"173            USER_FEEDBACK_STATE[user_id]["pending"] = True174            return reply, json.dumps(GLOBAL_CONVO_HISTORY[-5:], indent=2), ""175        else:176            USER_FEEDBACK_STATE[user_id] = {"emotion": None, "pending": False}177            return "How can I help you further?", json.dumps(GLOBAL_CONVO_HISTORY[-5:], indent=2), ""178 179    pred_emotion = get_emotion(user_input)180    support = random.choice(responses.get(pred_emotion, responses["neutral"]))181    try:182        suggestion = get_unique_solution(pred_emotion)183    except Exception:184        suggestion = get_unique_solution("neutral")185 186    reply = f"{support}\n\nHere's a suggestion for you: {suggestion}\nDid this help? (yes/no/skip)"187    GLOBAL_CONVO_HISTORY.append({188        "user_input": user_input,189        "emotion": pred_emotion,190        "bot_support": support,191        "bot_suggestion": suggestion,192        "feedback": ""193    })194    USER_FEEDBACK_STATE[user_id] = {"emotion": pred_emotion, "pending": True}195    return reply, json.dumps(GLOBAL_CONVO_HISTORY[-5:], indent=2), ""196 197# ---- Gradio interface ----198iface = gr.Interface(199    fn=emoti_chat,200    inputs=[201        gr.Audio(type="filepath", label="🎤 Speak your message"),202        gr.Textbox(lines=2, placeholder="Or type your message here...", label="💬 Type message"),203        gr.Textbox(lines=1, value="", visible=False)  # hidden, history state204    ],205    outputs=[206        gr.Textbox(label="EmotiBot Reply"),207        gr.Textbox(label="Hidden", visible=False)208    ],209    title="EmotiBot Connect",210    description="Talk to EmotiBot using your voice or by typing. Detects your emotion, gives dynamic suggestions, remembers your feedback, and keeps a conversation history! Type 'exit' to leave."211)212 213if __name__ == "__main__":214    iface.launch(debug=True) 215