app.py

import torch
import gradio as gr
import speech_recognition as sr
import time
import difflib
import random
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
from happytransformer import HappyTextToText, TTSettings

# Constants
MODEL_NAME = "prithivida/grammar_error_correcter_v1"
CSS = """
.gradio-container { max-width: 1400px !important; }
.header { text-align: center; padding: 2rem; background: linear-gradient(135deg, #3b82f6, #6366f1); color: white; border-radius: 15px; }
#container { height: 500px; width: 100%; background: #1a1a1a; border-radius: 10px; }
.diff-ins { color: #10b981; background: #d1fae5; padding: 2px 4px; border-radius: 4px; }
.diff-del { color: #ef4444; background: #fee2e2; padding: 2px 4px; border-radius: 4px; }
"""

# Three.js Template
THREEJS_TEMPLATE = """
<div id="container"></div>
<script async src="https://unpkg.com/[email protected]/dist/es-module-shims.js"></script>
<script type="importmap">
{
  "imports": {
    "three": "https://unpkg.com/[email protected]/build/three.module.js",
    "three/addons/": "https://unpkg.com/[email protected]/examples/jsm/"
  }
}
</script>

<script type="module">
import * as THREE from 'three';
import { OrbitControls } from 'three/addons/controls/OrbitControls.js';

class GrammarVisualizer {
  constructor() {
    this.initScene();
    this.addLights();
    this.createGrammarSphere();
    this.setupControls();
    this.animate();
  }

  initScene() {
    this.scene = new THREE.Scene();
    this.camera = new THREE.PerspectiveCamera(75, 500/400, 0.1, 1000);
    this.renderer = new THREE.WebGLRenderer({ antialias: true, alpha: true });
    document.getElementById('container').appendChild(this.renderer.domElement);
    this.renderer.setSize(500, 400);
    this.camera.position.z = 5;
  }

  addLights() {
    const ambient = new THREE.AmbientLight(0x404040);
    const directional = new THREE.DirectionalLight(0xffffff, 1);
    directional.position.set(5, 5, 5);
    this.scene.add(ambient, directional);
  }

  createGrammarSphere() {
    const geometry = new THREE.SphereGeometry(2, 32, 32);
    this.material = new THREE.MeshPhongMaterial({ 
      color: 0x3b82f6,
      transparent: true,
      opacity: 0.9
    });
    this.sphere = new THREE.Mesh(geometry, this.material);
    this.scene.add(this.sphere);
  }

  setupControls() {
    this.controls = new OrbitControls(this.camera, this.renderer.domElement);
    this.controls.enableDamping = true;
    this.controls.dampingFactor = 0.05;
  }

  animate() {
    requestAnimationFrame(() => this.animate());
    this.controls.update();
    this.renderer.render(this.scene, this.camera);
  }

  updateVisuals(score) {
    const hue = score / 100;
    this.material.color.setHSL(hue, 0.8, 0.5);
    this.sphere.rotation.x = (score / 50) * Math.PI;
    this.sphere.rotation.y = (score / 75) * Math.PI;
  }
}

let visualizer;
window.addEventListener('DOMContentLoaded', () => {
  visualizer = new GrammarVisualizer();
});

window.updateGrammarVisuals = (score) => {
  if(visualizer) visualizer.updateVisuals(score);
};
</script>
"""

# Load models
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_NAME)
happy_tt = HappyTextToText("T5", MODEL_NAME)

def create_diff_html(original, corrected):
    d = difflib.Differ()
    diff = d.compare(original.split(), corrected.split())    
    return " ".join([
        f'<span class="diff-ins">{p[2:]}</span> ' if p.startswith('+ ') else
        f'<span class="diff-del">{p[2:]}</span> ' if p.startswith('- ') else
        f'{p[2:]} ' for p in diff
    ])

def analyze_grammar(text):
    inputs = tokenizer.encode("gec: " + text, return_tensors="pt", max_length=256, truncation=True)
    with torch.no_grad():
        outputs = model.generate(inputs, max_length=256, num_beams=5)
    correction = tokenizer.decode(outputs[0], skip_special_tokens=True)
    
    args = TTSettings(num_beams=5, min_length=1)
    happy_correction = happy_tt.generate_text("gec: " + text, args=args).text
    
    final_correction = happy_correction if len(happy_correction) > len(correction) else correction
    changes = sum(1 for a, b in zip(text.split(), final_correction.split()) if a != b)
    score = max(0, 100 - (changes * 2))
    
    return {
        "original": text,
        "corrected": final_correction,
        "score": score,
        "diff_html": create_diff_html(text, final_correction)
    }

def process_input(audio_path, text):
    if audio_path:
        recognizer = sr.Recognizer()
        with sr.AudioFile(audio_path) as source:
            audio = recognizer.record(source)
        try:
            text = recognizer.recognize_google(audio)
        except sr.UnknownValueError:
            return "Could not understand audio", 0, "", ""
    
    if not text.strip():
        return "No input", 0, "", ""
    
    results = analyze_grammar(text)
    return [
        results['original'],
        results['score'],
        results['diff_html'],
        f"<script>window.updateGrammarVisuals({results['score']})</script>"
    ]

with gr.Blocks(css=CSS) as app:
    gr.Markdown("""
    <div class="header">
        <h1>🌍 3D Grammar Analyzer Pro</h1>
        <p>Interactive AI-Powered Writing Assistant with 3D Visualization</p>
    </div>
    """)
    
    with gr.Row():
        with gr.Column(scale=1):
            gr.Markdown("### Input Section")
            audio_input = gr.Audio(sources=["microphone"], type="filepath", label="🎤 Voice Input")
            text_input = gr.Textbox(lines=5, placeholder="📝 Type your text here...", label="Text Input")
            submit_btn = gr.Button("🚀 Analyze Text", variant="primary")
            
        with gr.Column(scale=2):
            gr.Markdown("### 3D Visualization")
            threejs = gr.HTML(THREEJS_TEMPLATE)
            
            with gr.Row():
                grammar_score = gr.Number(label="📊 Grammar Score", precision=0)
                score_gauge = gr.BarPlot(x=["Score"], y=[0], color="#3b82f6", height=150)
                
            diff_output = gr.HTML(label="📝 Text Corrections")
            hidden_trigger = gr.HTML(visible=False)

    gr.Markdown("### Example Sentences")
    gr.Examples(
        examples=[
            ["I is going to the park yesterday."],
            ["Their happy about there test results."],
            ["She dont like apples, but she like bananas."]
        ],
        inputs=[text_input],
        outputs=[text_input, grammar_score, diff_output, hidden_trigger],
        fn=process_input,
        cache_examples=True
    )
    
    submit_btn.click(
        fn=process_input,
        inputs=[audio_input, text_input],
        outputs=[text_input, grammar_score, diff_output, hidden_trigger]
    )
    
    text_input.change(
        lambda x: analyze_grammar(x)["score"] if x else 0,
        inputs=text_input,
        outputs=grammar_score
    )

if __name__ == "__main__":
    app.launch(
        server_name="0.0.0.0",
        server_port=7860,
        share=True,
        favicon_path="https://raw.githubusercontent.com/gradio-app/gradio/main/guides/assets/logo.png"
    )