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 - from educationplanner.org
learning_style_questions = [
    "When you are learning something new, you prefer to:",
    "When you are at home, you like to:",
    "When you spell a word, you remember it by:",
    "When you read, you:",
    "When you write, you:",
    "When you listen to music, you:",
    "When you work at solving a problem, you:",
    "When you give someone directions, you:",
    "When you are concentrating, you:",
    "When you meet someone new, you remember them by:"
]

learning_style_answers = [
    ["Watch a demonstration", "Listen to instructions", "Read about it"],
    ["Watch TV or play games", "Listen to music or talk", "Read books or write"],
    ["Seeing the word", "Hearing the word", "Writing the word"],
    ["See the characters in your mind", "Hear the words in your mind", "Focus on the meaning of the words"],
    ["Prefer diagrams or charts", "Prefer to discuss ideas", "Prefer to write detailed explanations"],
    ["Enjoy the sound and rhythm", "Remember lyrics easily", "Analyze the meaning"],
    ["Visualize the problem", "Talk it through", "Write out steps"],
    ["Use maps or draw it", "Give verbal directions", "Write down directions"],
    ["Create mental pictures", "Repeat things aloud", "Take notes or read quietly"],
    ["Facial features", "Voice or name", "Spelling or written form"]
]

style_count_map = {
    0: 'visual',
    1: 'auditory',
    2: 'reading/writing'
}

# Quiz logic to analyze learning style
def learning_style_quiz(*answers):
    scores = {'visual': 0, 'auditory': 0, 'reading/writing': 0}
    for ans in answers:
        scores[ans] += 1
    best = max(scores, key=scores.get)
    return best.capitalize()

# PanoramaEd categories and multiple choice questions
get_to_know_categories = {
    "All About Me": [
        ("What’s your favorite way to spend a day off?", []),
        ("If you could only eat one food for the rest of your life, what would it be?", []),
        ("Do you have any pets? If so, what are their names?", []),
        ("If you could travel anywhere in the world, where would you go?", []),
        ("What’s your favorite holiday or tradition?", [])
    ],
    "Hopes and Dreams": [
        ("What do you want to be when you grow up?", []),
        ("What’s something you hope to achieve this year?", []),
        ("If you could change the world in one way, what would you do?", []),
        ("What are you most proud of?", []),
        ("What’s a big dream you have for your future?", [])
    ],
    "School Life": [
        ("What’s your favorite subject in school?", []),
        ("What’s something that makes learning easier for you?", []),
        ("Do you prefer working alone or in groups?", []),
        ("What helps you feel confident in class?", []),
        ("What’s something you’re good at in school?", [])
    ],
    "Relationships": [
        ("Who do you look up to and why?", []),
        ("Who is someone that makes you feel safe and supported?", []),
        ("Do you have a best friend? What do you like to do together?", []),
        ("What’s one thing you wish people knew about you?", []),
        ("What’s something kind you’ve done for someone else?", [])
    ]
}

# Save all answers into profile
def save_profile(file, *inputs):
    quiz_answers = inputs[:len(learning_style_questions)]
    about_me = inputs[len(learning_style_questions)]
    blog_opt_in = inputs[len(learning_style_questions)+1]
    blog_text = inputs[len(learning_style_questions)+2]
    category_answers = inputs[len(learning_style_questions)+3:]

    df = parse_transcript(file)
    transcript_info = extract_transcript_info(df)
    learning_type = learning_style_quiz(*quiz_answers)

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

    question_texts = [q for cat in get_to_know_categories.values() for q, _ in cat]
    responses = dict(zip(question_texts, category_answers))

    profile = {
        "transcript": df.to_dict(orient='records'),
        "transcript_info": transcript_info,
        "learning_style": learning_type,
        "about_me": about_me,
        "get_to_know_answers": responses,
        "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")
        quiz_components = []
        for i, (question, options) in enumerate(zip(learning_style_questions, learning_style_answers)):
            quiz_components.append(gr.Radio(
                choices=options,
                label=f"{i+1}. {question}"
            ))

    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)

    category_inputs = []
    for category, questions in get_to_know_categories.items():
        gr.Markdown(f"### 📘 {category}")
        for q_text, _ in questions:
            category_inputs.append(gr.Textbox(label=q_text))

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

    submit.click(fn=save_profile,
                 inputs=[file, *quiz_components, about_me, blog_opt_in, blog_text, *category_inputs],
                 outputs=[output])

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