app.py

import gradio as gr
import pandas as pd
import PyPDF2
import json
import re

# Parse uploaded transcript file
def parse_transcript(file):
    if file.name.endswith('.csv'):
        df = pd.read_csv(file.name)
    elif file.name.endswith(('.xls', '.xlsx')):
        df = pd.read_excel(file.name)
    elif file.name.endswith('.pdf'):
        reader = PyPDF2.PdfReader(file)
        text = ""
        for page in reader.pages:
            text += page.extract_text() or ""
        df = pd.DataFrame({'Transcript_Text': [text]})
    else:
        raise ValueError("Unsupported file format. Use .csv, .xlsx, or .pdf")
    return df

# Extract student info
def extract_transcript_info(df):
    transcript_text = df['Transcript_Text'].iloc[0] if 'Transcript_Text' in df.columns else ''
    info = {}
    gpa_match = re.search(r'(GPA|Grade Point Average)[^\d]*(\d+\.\d+)', transcript_text, re.IGNORECASE)
    if gpa_match:
        info['GPA'] = gpa_match.group(2)
    grade_match = re.search(r'Grade:?\s*(\d{1,2})', transcript_text, re.IGNORECASE)
    if grade_match:
        info['Grade_Level'] = grade_match.group(1)
    courses = re.findall(r'(?i)\b([A-Z][a-zA-Z\s&/]+)\s+(\d{1,3})\b', transcript_text)
    if courses:
        info['Courses'] = list(set([c[0].strip() for c in courses]))
    return info

# Learning style questions
def learning_style_quiz(q1, q2, q3):
    scores = {'visual': 0, 'auditory': 0, 'reading/writing': 0, 'kinesthetic': 0}
    mapping = [q1, q2, q3]
    for answer in mapping:
        scores[answer] += 1
    best = max(scores, key=scores.get)
    return best.capitalize()

# Save all answers into profile
def save_profile(file, q1, q2, q3, about_me, blog_text, blog_opt_in):
    df = parse_transcript(file)
    transcript_info = extract_transcript_info(df)
    learning_type = learning_style_quiz(q1, q2, q3)

    if not blog_opt_in and blog_text.strip() == "":
        blog_text = "[User chose to skip this section]"

    profile = {
        "transcript": df.to_dict(orient='records'),
        "transcript_info": transcript_info,
        "learning_style": learning_type,
        "about_me": about_me,
        "blog": blog_text
    }

    with open("student_profile.json", "w") as f:
        json.dump(profile, f, indent=4)

    return f"✅ Profile saved! Your learning style is: {learning_type}"

# Build Gradio UI
with gr.Blocks() as demo:
    gr.Markdown("## 🎓 Personalized AI Student Assistant")

    with gr.Row():
        file = gr.File(label="📄 Upload Your Transcript (.csv, .xlsx, .pdf)")

    with gr.Column():
        gr.Markdown("### 🧠 Learning Style Discovery")
        q1 = gr.Radio(["visual", "auditory", "reading/writing", "kinesthetic"], label="1. How do you prefer to learn new topics?")
        q2 = gr.Radio(["visual", "auditory", "reading/writing", "kinesthetic"], label="2. How do you remember lists best?")
        q3 = gr.Radio(["visual", "auditory", "reading/writing", "kinesthetic"], label="3. Favorite way to study?")

    with gr.Column():
        gr.Markdown("### ❤️ About You")
        about_me = gr.Textbox(lines=6, label="Answer a few questions: \n1. What’s a fun fact about you? \n2. Favorite music/artist? \n3. Your dream job?")

        blog_opt_in = gr.Checkbox(label="I want to write a personal blog for better personalization")
        blog_text = gr.Textbox(lines=5, label="✍️ Optional: Write a mini blog about your life", visible=True)

    submit = gr.Button("📥 Save My Profile")
    output = gr.Textbox(label="Status")

    submit.click(fn=save_profile,
                 inputs=[file, q1, q2, q3, about_me, blog_text, blog_opt_in],
                 outputs=[output])

if __name__ == '__main__':
    demo.launch()