# -*- coding: utf-8 -*-
"""Untitled4.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1352Z_3Tsa5_YFTfI-jWhZpSJ_k4yHSm3
"""

# pip install gradio transformers torch

# !pip install gTTS

import gradio as gr
from transformers import pipeline, TextGenerationPipeline, AutoModelForCausalLM, AutoTokenizer
from transformers import AutoTokenizer, AutoModelForCausalLM, TextGenerationPipeline
import torch
from gtts import gTTS
import tempfile

sentiment_pipeline = pipeline("sentiment-analysis")
summarizer_pipeline = pipeline("summarization")


# Sentiment Analysis
def analyze_sentiment(text):
    result = sentiment_pipeline(text)[0]
    return result["label"], round(result["score"], 3)


# Summarization
def summarize(text):
    summary = summarizer_pipeline(text, max_length=60, min_length=15, do_sample=False)
    return summary[0]["summary_text"]


def text_to_speech(text):
    tts = gTTS(text)
    with tempfile.NamedTemporaryFile(suffix=".mp3", delete=False) as fp:
        tts.save(fp.name)
        return fp.name


# Gradio UI
with gr.Blocks(title="TrailTrek AI Assistant",theme="soft") as demo:
    gr.Markdown("## 🧠 TrailTrek Gears Co - Multi-Task AI Demo")

    with gr.Tab("📊 Sentiment Analysis"):
        with gr.Row():
            text_input = gr.Textbox(label="Enter text")
            sentiment_output = gr.Text(label="Sentiment")
            confidence_output = gr.Number(label="Confidence")
        analyze_btn = gr.Button("Analyze")
        analyze_btn.click(analyze_sentiment, inputs=[text_input], outputs=[sentiment_output, confidence_output])


    with gr.Tab("📄 Summarization"):
        input_text = gr.Textbox(lines=8, label="Enter a long text")
        output_summary = gr.Text(label="Summary")
        summarize_btn = gr.Button("Summarize")
        summarize_btn.click(summarize, inputs=[input_text], outputs=[output_summary])


    with gr.Tab("🗣️ Text-to-Speech"):
        tts_input = gr.Textbox(label="Enter text to speak")
        tts_output = gr.Audio(label="Generated Speech", type="filepath")
        tts_btn = gr.Button("Convert to Speech")
        tts_btn.click(text_to_speech, inputs=[tts_input], outputs=[tts_output])



demo.launch()