app.py

import gradio as gr
import json
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
import logging
from pathlib import Path
import tempfile
import time

# Setup logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

class SyllabusFormatter:
    def __init__(self, model_name="microsoft/Phi-3-mini-4k-instruct"):
        self.model_name = model_name
        self.tokenizer = None
        self.model = None
        self.pipe = None
        self.processed_count = 0
        self.total_count = 0
        
    def setup_model(self):
        """Download and setup the Phi model"""
        try:
            # Load tokenizer
            self.tokenizer = AutoTokenizer.from_pretrained(
                self.model_name,
                trust_remote_code=True
            )
            
            # Load model with 8-bit quantization for efficiency
            self.model = AutoModelForCausalLM.from_pretrained(
                self.model_name,
                torch_dtype=torch.float16,
                device_map="auto",
                trust_remote_code=True,
                load_in_8bit=True
            )
            
            # Create pipeline
            self.pipe = pipeline(
                "text-generation",
                model=self.model,
                tokenizer=self.tokenizer,
                max_new_tokens=2048,
                temperature=0.1,
                do_sample=True,
                top_p=0.9,
                repetition_penalty=1.1
            )
            
            logger.info("Model setup complete!")
            return True
            
        except Exception as e:
            logger.error(f"Error setting up model: {str(e)}")
            return False

    def create_formatting_prompt(self, unit_content: str, unit_name: str, subject_name: str = "") -> str:
        """Create a very clear, focused prompt for formatting syllabus content"""
        prompt = f"""<|system|>You are a professional academic syllabus formatter. Your ONLY job is to take badly formatted syllabus content and make it beautifully organized and readable.



RULES:

1. PRESERVE every single word, topic, and concept from the original

2. NEVER add explanations, examples, or new content

3. ONLY restructure and format the existing text

4. Use clear headings, bullet points, and logical grouping

5. Separate different topics with proper spacing

6. Make it scannable and easy to read



FORMAT STYLE:

- Use main topic headings with proper capitalization

- Group related subtopics under main topics

- Use bullet points (•) for lists of concepts

- Use sub-bullets (◦) for details under main bullets

- Separate major sections with line breaks

- Keep technical terms exactly as written<|end|>



<|user|>Subject: {subject_name}

Unit: {unit_name}



Original content (poorly formatted):

{unit_content}



Task: Reformat this content to be beautifully organized and readable. Do NOT add any new information - only restructure what's already there. Make it professional and easy to scan.<|end|>



<|assistant|>"""
        return prompt
    
    def format_unit_content(self, unit_content: str, unit_name: str, subject_name: str = "", progress=None) -> str:
        """Format a single unit's content using the AI model"""
        try:
            # Create prompt
            prompt = self.create_formatting_prompt(unit_content, unit_name, subject_name)
            
            # Generate formatted content
            response = self.pipe(prompt)
            
            # Extract formatted content
            generated_text = response[0]['generated_text']
            assistant_start = generated_text.find("<|assistant|>")
            if assistant_start != -1:
                formatted_content = generated_text[assistant_start + len("<|assistant|>"):].strip()
            else:
                formatted_content = generated_text.strip()
            
            # Clean up and validate
            formatted_content = self.clean_generated_content(formatted_content)
            if not self.validate_formatted_content(unit_content, formatted_content):
                return unit_content
                
            return formatted_content
            
        except Exception as e:
            logger.error(f"Error formatting content: {str(e)}")
            return unit_content
            
    def validate_formatted_content(self, original: str, formatted: str) -> bool:
        """Validate that formatted content preserves all important information"""
        # Basic validation
        if len(formatted) < len(original) * 0.4:
            return False
        return True
            
    def clean_generated_content(self, content: str) -> str:
        """Clean up generated content"""
        # Remove special tokens
        for token in ["<|system|>", "<|user|>", "<|assistant|>"]:
            content = content.replace(token, "")
            
        # Clean up extra whitespace
        content = "\n".join(line.strip() for line in content.split("\n") if line.strip())
        
        return content

    def process_syllabus_file(self, syllabus_data: dict, progress=gr.Progress()) -> dict:
        """Process the entire syllabus file with progress updates"""
        try:
            # Count total units
            total_units = 0
            processed = 0
            
            def count_units(data):
                count = 0
                if isinstance(data, dict):
                    for value in data.values():
                        if isinstance(value, dict):
                            count += count_units(value)
                        elif isinstance(value, str) and "Unit" in str(value):
                            count += 1
                return count
            
            total_units = count_units(syllabus_data.get("syllabus", {}))
            logger.info(f"Total units to process: {total_units}")
            
            # Process each branch
            for branch_name, branch_data in syllabus_data.get("syllabus", {}).items():
                if not isinstance(branch_data, dict):
                    continue
                    
                # Process each semester
                for sem_name, sem_data in branch_data.items():
                    if not isinstance(sem_data, dict):
                        continue
                        
                    # Process each subject
                    for subject_name, subject_data in sem_data.items():
                        if not isinstance(subject_data, dict) or "content" not in subject_data:
                            continue
                            
                        content = subject_data["content"]
                        if not isinstance(content, dict):
                            continue
                        
                        # Format each unit
                        for unit_name, unit_content in content.items():
                            if not unit_name.startswith("Unit") or not isinstance(unit_content, str):
                                continue
                            
                            processed += 1
                            progress(processed / total_units, desc=f"Processing {unit_name} in {subject_name}...")
                            
                            # Format the unit content
                            formatted_content = self.format_unit_content(
                                unit_content, 
                                unit_name, 
                                subject_name
                            )
                            
                            # Update the content
                            syllabus_data["syllabus"][branch_name][sem_name][subject_name]["content"][unit_name] = formatted_content
            
            # Add formatting metadata
            if "metadata" not in syllabus_data:
                syllabus_data["metadata"] = {}
                
            syllabus_data["metadata"].update({
                "lastFormatted": time.strftime("%Y-%m-%dT%H:%M:%SZ"),
                "formattingModel": "Phi-3 Mini",
                "unitsProcessed": processed,
                "version": "1.0"
            })
            
            return syllabus_data
            
        except Exception as e:
            logger.error(f"Error processing syllabus: {str(e)}")
            raise gr.Error(f"Error processing syllabus: {str(e)}")

# Initialize the formatter
formatter = None

def setup_formatter():
    global formatter
    if formatter is None:
        formatter = SyllabusFormatter()
        return formatter.setup_model()
    return True

def process_file(file):
    """Process the uploaded syllabus file"""
    try:
        # Setup formatter if needed
        if not setup_formatter():
            raise gr.Error("Failed to setup the formatting model. Please try again.")
        
        # Read and parse JSON
        content = file.read()
        syllabus_data = json.loads(content)
        
        # Process syllabus
        formatted_data = formatter.process_syllabus_file(syllabus_data)
        
        # Save to temporary file
        with tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False) as tmp:
            json.dump(formatted_data, tmp, indent=2)
            return tmp.name
            
    except json.JSONDecodeError:
        raise gr.Error("Invalid JSON file. Please check your syllabus file format.")
    except Exception as e:
        raise gr.Error(f"Error processing file: {str(e)}")

# Custom theme
theme = gr.themes.Soft(
    primary_hue="indigo",
    secondary_hue="blue",
).set(
    body_background_fill="#fafafa",
    body_background_fill_dark="#1a1a1a",
    button_primary_background_fill="*primary_500",
    button_primary_background_fill_hover="*primary_600"
)

# Gradio interface
title = "📚 Syllabus Formatter"
description = """

Transform your syllabus into a beautifully formatted, easy-to-read document using AI.



### Features:

- Preserves all original content

- Improves readability and organization

- Creates logical grouping and sections

- Adds professional formatting



Simply upload your JSON syllabus file and get a formatted version back!

"""

css = """

.feedback {

    margin-top: 20px;

    padding: 10px;

    border-radius: 8px;

    background-color: #f0f9ff;

    border: 1px solid #bae6fd;

}

.dark .feedback {

    background-color: #082f49;

    border-color: #075985;

}

"""

with gr.Blocks(theme=theme, css=css) as iface:
    gr.Markdown(f"# {title}")
    gr.Markdown(description)
    
    with gr.Row():
        with gr.Column():
            file_input = gr.File(
                label="Upload Syllabus JSON",
                file_types=[".json"],
                file_count="single"
            )
            process_btn = gr.Button("🪄 Format Syllabus", variant="primary")
            
    output_file = gr.File(
        label="Download Formatted Syllabus",
        file_count="single",
        type="file",
        interactive=False
    )
    
    with gr.Row():
        feedback = gr.Markdown(
            value="Upload a JSON syllabus file to begin...",
            elem_classes=["feedback"]
        )
    
    def update_feedback(file):
        return "Processing your syllabus... This may take a few minutes depending on the size."
        
    # Setup click event
    process_btn.click(
        fn=update_feedback,
        inputs=[file_input],
        outputs=[feedback],
        queue=False
    ).then(
        fn=process_file,
        inputs=[file_input],
        outputs=[output_file]
    ).success(
        fn=lambda: "✨ Syllabus formatting complete! You can now download the formatted file.",
        outputs=[feedback]
    )

    gr.Markdown("""

    ### 📝 Notes:

    - The formatter preserves all original content while improving organization

    - Processing time depends on the size of your syllabus

    - For large files, please be patient as the AI processes each section

    

    Made with ❤️ using Microsoft's Phi-3 Mini model

    """)

# Launch in dev mode
if __name__ == "__main__":
    iface.launch()