app.py

"""
GradeScope — CGPA Predictor · Flask Backend
============================================
Architecture:
  - 5-page HTML frontend served from /
  - Single POST endpoint /api/predict

Encoding reference (matches training data exactly):
  Career alignment / Participation / Peer behaviour / Interest / Difficulty (student-perceived)
      Very Poor/Very Low = 1  |  Poor/Low = 2  |  Good/High = 4  |  Very Strong/Very High = 5
      (scale jumps 2→4, skipping 3 — matches original survey encoding)

  Seating:   Front = 1  |  Random = 3  |  Back = 5
  Teacher:   Poor+Strict = 1  |  Good+Strict = 3  |  Poor+Friendly = 4  |  Good+Friendly = 6
  Attendance: Below 60% = 1  |  60-75% = 2  |  75-90% = 3  |  91-100% = 4

  ec_difference  = extracurricular_encoded (1/2/4/5) − absolute_difficulty_from_CSV (1-4)
  diff_alignment = career_alignment_encoded − student_perceived_difficulty
  relative_diff  = predicted_sgpa − absolute_difficulty   [computed inline]
"""

from flask import Flask, render_template, request, jsonify
import pickle, json, traceback
import numpy as np
import os

app = Flask(__name__)
BASE = os.path.dirname(__file__)


# ── Load trained models ───────────────────────────────────────────
def _load(name):
    with open(os.path.join(BASE, 'models', name), 'rb') as f:
        return pickle.load(f)


rf_model    = _load('rf_model.pkl')
gb_model    = _load('gb_model.pkl')
et_model    = _load('et_model.pkl')
ridge_model = _load('ridge_model.pkl')
imputer     = _load('imputer.pkl')
scaler      = _load('scaler.pkl')

with open(os.path.join(BASE, 'models', 'model_stats.json')) as f:
    STATS = json.load(f)

FEATURE_COLS = STATS['feature_cols']
W = STATS['ensemble_weights']          # {rf, gb, et, ridge}
W_TOTAL = sum(W.values())


# ── Subject metadata ──────────────────────────────────────────────
SUBJECTS_DATA = {
    # Institute Core
    "Engineering Drawing and Sketching":    {"credits": 1, "abs_diff": 3},
    "Basic Economics":                      {"credits": 2, "abs_diff": 1},
    "English Communication Skills":         {"credits": 2, "abs_diff": 1},
    "Introduction to Mechanical Systems":   {"credits": 2, "abs_diff": 2},
    "Modern Physics":                       {"credits": 3, "abs_diff": 3},
    "Mathematics II":                       {"credits": 4, "abs_diff": 4},
    "Classical Physics":                    {"credits": 3, "abs_diff": 2},
    "Environmental Science":                {"credits": 2, "abs_diff": 1},
    "Programming with Python":              {"credits": 2, "abs_diff": 2},
    "Engineering Chemistry":                {"credits": 3, "abs_diff": 2},
    "Basics of Electronics Engineering":    {"credits": 2, "abs_diff": 2},
    "Basics of Electrical Engineering":     {"credits": 2, "abs_diff": 3},
    "Mathematics I":                        {"credits": 4, "abs_diff": 3},
    # MECH
    "Applied Probability and Statistics":   {"credits": 3, "abs_diff": 2},
    "Casting Welding and Forming":          {"credits": 3, "abs_diff": 2},
    "Problem Solving using C":              {"credits": 2, "abs_diff": 2},
    "Discrete Mathematics":                 {"credits": 3, "abs_diff": 3},
    "Engineering Mechanics":                {"credits": 3, "abs_diff": 2},
    "Engineering Thermodynamics":           {"credits": 4, "abs_diff": 4},
    "Kinematics of Machines":               {"credits": 3, "abs_diff": 2},
    "Basics of Management":                 {"credits": 3, "abs_diff": 1},
    "Fluid Mechanics":                      {"credits": 4, "abs_diff": 3},
    "Materials Science and Engineering":    {"credits": 3, "abs_diff": 2},
    "Mechanical Measurements and Metrology":{"credits": 3, "abs_diff": 2},
    "Solid Mechanics":                      {"credits": 4, "abs_diff": 3},
    "Operations Planning and Control":      {"credits": 3, "abs_diff": 2},
    "Applied machine learning":             {"credits": 3, "abs_diff": 3},
    "Mechatronics in Manufacturing":        {"credits": 3, "abs_diff": 3},
    "Dynamics of Machines":                 {"credits": 4, "abs_diff": 4},
    "Heat Transfer":                        {"credits": 4, "abs_diff": 4},
    "I C Engines":                          {"credits": 3, "abs_diff": 2},
    "Science of Machining":                 {"credits": 4, "abs_diff": 3},
    "Introduction to Standardization":      {"credits": 1, "abs_diff": 1},
    "Industrial Engineering":               {"credits": 2, "abs_diff": 2},
    "Operations Research":                  {"credits": 3, "abs_diff": 4},
    "Control System Engineering":           {"credits": 4, "abs_diff": 4},
    "CAD & CAM":                            {"credits": 3, "abs_diff": 2},
    "Design of Machine Elements":           {"credits": 4, "abs_diff": 3},
    "Fluid and Turbo Machines":             {"credits": 4, "abs_diff": 3},
    "Refrigeration and Air Conditioning":   {"credits": 3, "abs_diff": 3},
    "Product Design and Development":       {"credits": 3, "abs_diff": 2},
    "Supply Chain Management":              {"credits": 3, "abs_diff": 2},
    "Advanced Machining & Additive Manufacturing":{"credits": 3, "abs_diff": 3},
    "Design of Mechanical Systems":         {"credits": 4, "abs_diff": 4},
    "Automobile Engineering":               {"credits": 3, "abs_diff": 2},
    "Industry 4.0":                         {"credits": 3, "abs_diff": 2},
    "Mechanics of Composites":              {"credits": 3, "abs_diff": 4},
    "Metal Forming":                        {"credits": 3, "abs_diff": 3},
    "Project Management":                   {"credits": 3, "abs_diff": 1},
    "Air Conditioning System Design":       {"credits": 3, "abs_diff": 3},
    "Artificial Intelligence for Robotic systems":{"credits": 3, "abs_diff": 4},
    "Lean Six Sigma":                       {"credits": 3, "abs_diff": 2},
    "Tool Engineering":                     {"credits": 3, "abs_diff": 3},
    "Welding Engineering & Technology":     {"credits": 3, "abs_diff": 2},
    # META
    "Introduction to Engineering Materials":{"credits": 3, "abs_diff": 1},
    "Fuels, Furnaces and Refractories":     {"credits": 4, "abs_diff": 2},
    "Introduction to Physical Metallurgy":  {"credits": 4, "abs_diff": 3},
    "Mineral Processing":                   {"credits": 3, "abs_diff": 2},
    "Electrometallurgy & Corrosion":        {"credits": 4, "abs_diff": 3},
    "Foundry Technology":                   {"credits": 4, "abs_diff": 2},
    "Introduction to Extractive Metallurgy":{"credits": 4, "abs_diff": 3},
    "Mechanical Behaviour and Testing of Metals":{"credits": 4, "abs_diff": 3},
    "Metallurgical Thermodynamics and Kinetics":{"credits": 4, "abs_diff": 4},
    "Computer Aided Design and Computer Aided Manufacturing (Program Linked EAS)":{"credits": 2, "abs_diff": 2},
    "Basics of Transport Phenomena":        {"credits": 4, "abs_diff": 4},
    "Iron Making":                          {"credits": 4, "abs_diff": 3},
    "Mechanical Working of Metals":         {"credits": 4, "abs_diff": 3},
    "NDT and Evaluation":                   {"credits": 3, "abs_diff": 2},
    "Phase Transformations":                {"credits": 4, "abs_diff": 4},
    "Introduction to Nanomaterials and Technology":{"credits": 3, "abs_diff": 3},
    "Materials Characterization":           {"credits": 3, "abs_diff": 3},
    "Powder Metallurgy":                    {"credits": 3, "abs_diff": 2},
    "Principles of Heat Treatment":         {"credits": 4, "abs_diff": 3},
    "Production of Sponge Iron and Ferroalloys":{"credits": 3, "abs_diff": 2},
    "Design and selection of materials":    {"credits": 3, "abs_diff": 3},
    "High Temperature Corrosion":           {"credits": 3, "abs_diff": 3},
    "Composite Materials":                  {"credits": 4, "abs_diff": 3},
    "Joining of Materials":                 {"credits": 4, "abs_diff": 2},
    "Non-Ferrous Extractive Metallurgy":    {"credits": 4, "abs_diff": 3},
    "Steel Making":                         {"credits": 4, "abs_diff": 3},
    "Light Metals and Alloys":              {"credits": 3, "abs_diff": 2},
    "Welding Metallurgy":                   {"credits": 3, "abs_diff": 3},
    "Fracture & Failure":                   {"credits": 3, "abs_diff": 4},
    "Physical Metallurgy of Alloy Steels":  {"credits": 3, "abs_diff": 4},
    "Surface Engineering":                  {"credits": 3, "abs_diff": 2},
    "Automotive and Aerospace Material":    {"credits": 3, "abs_diff": 2},
    "Metallurgy of Additive Manufacturing": {"credits": 3, "abs_diff": 3},
    "Secondary Steel Making":               {"credits": 3, "abs_diff": 3},
    "Electronic and Magnetic Materials":    {"credits": 3, "abs_diff": 4},
    # EE
    "Power Generation Sources":             {"credits": 4, "abs_diff": 1},
    "Network Theory":                       {"credits": 4, "abs_diff": 3},
    "Measurements and Actuators":           {"credits": 4, "abs_diff": 2},
    "Advanced Programming for Design":      {"credits": 3, "abs_diff": 2},
    "Analysis & Design of Digital Logic Circuits":{"credits": 4, "abs_diff": 2},
    "Electrical Machines-I":                {"credits": 4, "abs_diff": 3},
    "Integrated Electronics":               {"credits": 4, "abs_diff": 3},
    "Network, Systems and Signals":         {"credits": 4, "abs_diff": 4},
    "Advanced Engineering Mathematics":     {"credits": 4, "abs_diff": 4},
    "Control Systems Engineering":          {"credits": 3, "abs_diff": 3},
    "Electrical Machines-II":              {"credits": 4, "abs_diff": 4},
    "Microprocessor and Microcontroller":   {"credits": 4, "abs_diff": 3},
    "Power Electronics-I":                  {"credits": 4, "abs_diff": 3},
    "Transmission and Distribution Systems":{"credits": 4, "abs_diff": 2},
    "Principles of Communication Engineering":{"credits": 4, "abs_diff": 3},
    "Modern Control Engineering":           {"credits": 4, "abs_diff": 4},
    "Power Electronics-II":                {"credits": 4, "abs_diff": 4},
    "Power System Switchgear and Protection":{"credits": 4, "abs_diff": 3},
    "Advanced Power Transmission":          {"credits": 4, "abs_diff": 3},
    "Digital Signal Processing and Applications":{"credits": 4, "abs_diff": 4},
    "Electric Drives and Control":          {"credits": 4, "abs_diff": 4},
    "Industrial Machine Learning":          {"credits": 4, "abs_diff": 3},
    "Operation and Control of Power Systems":{"credits": 4, "abs_diff": 4},
    "Computer Aided Power System Analysis": {"credits": 4, "abs_diff": 3},
    "Computer Architecture and Organization":{"credits": 4, "abs_diff": 3},
    "Optimal Control Systems":              {"credits": 4, "abs_diff": 4},
    "Power System Restructuring, Deregulation and Economics":{"credits": 4, "abs_diff": 2},
    # ECE
    "Electronic Measurement & Instrumentation":{"credits": 3, "abs_diff": 2},
    "Electronics Devices and Circuits":     {"credits": 3, "abs_diff": 3},
    "Signals and Systems":                  {"credits": 3, "abs_diff": 4},
    "Analog Communication (M1)":            {"credits": 3, "abs_diff": 3},
    "Data Structures and Algorithms":       {"credits": 3, "abs_diff": 3},
    "Digital Logic Design (M2)":            {"credits": 3, "abs_diff": 2},
    "Electro-Magnetic Field Theory":        {"credits": 3, "abs_diff": 4},
    "Linear Integrated Circuits":           {"credits": 3, "abs_diff": 3},
    "Operating System Concepts":            {"credits": 3, "abs_diff": 2},
    "Technical Documentation":              {"credits": 1, "abs_diff": 1},
    "Analog CMOS IC":                       {"credits": 3, "abs_diff": 4},
    "Computer Architecture":                {"credits": 3, "abs_diff": 3},
    "Digital Communication Systems":        {"credits": 3, "abs_diff": 3},
    "Digital Signal Processing":            {"credits": 3, "abs_diff": 4},
    "Microwave Engineering":                {"credits": 3, "abs_diff": 3},
    "VLSI Testing & Testability":           {"credits": 3, "abs_diff": 3},
    "Antenna & Wave Propagation":           {"credits": 3, "abs_diff": 3},
    "Digital CMOS IC":                      {"credits": 3, "abs_diff": 3},
    "Embedded Systems":                     {"credits": 3, "abs_diff": 3},
    "Microprocessors":                      {"credits": 3, "abs_diff": 3},
    "Management Principles for Engineers":  {"credits": 3, "abs_diff": 1},
    "Computer & Network Security":          {"credits": 3, "abs_diff": 2},
    "Optical Communication Systems":        {"credits": 3, "abs_diff": 3},
    "Satellite & Radar Engineering":        {"credits": 3, "abs_diff": 3},
    "Wireless and 5G Communication":        {"credits": 3, "abs_diff": 3},
    "Neural Networks & Fuzzy Logic":        {"credits": 3, "abs_diff": 3},
    "Advanced Digital Communication Systems":{"credits": 3, "abs_diff": 4},
    "Advanced Semiconductor Devices":       {"credits": 3, "abs_diff": 4},
    "Advanced Optical Communication Systems":{"credits": 3, "abs_diff": 4},
    "VLSI signal processing architectures": {"credits": 3, "abs_diff": 4},
    "CAD Algorithms for Synthesis of VLSI System":{"credits": 3, "abs_diff": 4},
    "Artificial Intelligence and Expert Systems":{"credits": 3, "abs_diff": 3},
    # CIVIL
    "Surveying":                            {"credits": 3, "abs_diff": 2},
    "Engineering Geology":                  {"credits": 3, "abs_diff": 2},
    "Mechanics of Solids":                  {"credits": 4, "abs_diff": 3},
    "Advanced Surveying":                   {"credits": 2, "abs_diff": 3},
    "Building Technology":                  {"credits": 2, "abs_diff": 2},
    "Introduction to Satellite Based Positioning":{"credits": 2, "abs_diff": 2},
    "Construction Materials":               {"credits": 3, "abs_diff": 1},
    "Structural Analysis":                  {"credits": 4, "abs_diff": 4},
    "Numerical Methods & Optimization":     {"credits": 3, "abs_diff": 3},
    "Building Planning & Services":         {"credits": 2, "abs_diff": 2},
    "Design of RCC Structures":             {"credits": 3, "abs_diff": 3},
    "Environmental Engineering - I":        {"credits": 3, "abs_diff": 2},
    "Geotechnical Engineering - I":         {"credits": 4, "abs_diff": 3},
    "Hydraulic Engineering":                {"credits": 4, "abs_diff": 3},
    "Transportation Engineering - I":       {"credits": 3, "abs_diff": 2},
    "Design of RCC Systems":               {"credits": 4, "abs_diff": 4},
    "Design of Steel Structures":           {"credits": 4, "abs_diff": 4},
    "Environmental Engineering-II":         {"credits": 3, "abs_diff": 3},
    "Geotechnical Engineering-II":          {"credits": 4, "abs_diff": 4},
    "Hydrology":                            {"credits": 4, "abs_diff": 2},
    "Transportation Engineering - II":      {"credits": 3, "abs_diff": 3},
    "Soft Skills and Personality Development":{"credits": 2, "abs_diff": 1},
    "Construction Project Management":      {"credits": 3, "abs_diff": 2},
    "Estimation and Costing":              {"credits": 4, "abs_diff": 2},
    "Water Resources Engineering":          {"credits": 3, "abs_diff": 3},
    "Numerical Methods":                    {"credits": 3, "abs_diff": 3},
    "Design of Hydraulic Structures":       {"credits": 3, "abs_diff": 4},
    "Dynamics of Structures":               {"credits": 3, "abs_diff": 4},
    "Earth Dams":                           {"credits": 3, "abs_diff": 3},
    "Finite Element Method":                {"credits": 3, "abs_diff": 4},
    "Landfill Engineering":                 {"credits": 3, "abs_diff": 2},
    "Prestressed Concrete":                 {"credits": 3, "abs_diff": 4},
    "Soil Dynamics":                        {"credits": 3, "abs_diff": 4},
    "Solid Waste Management":              {"credits": 3, "abs_diff": 2},
    # CSE
    "Technical Writing":                    {"credits": 1, "abs_diff": 1},
    "Data Communications":                  {"credits": 3, "abs_diff": 2},
    "Design and Analysis of Algorithms":    {"credits": 3, "abs_diff": 4},
    "Digital Circuits and Microprocessors": {"credits": 3, "abs_diff": 3},
    "MATHEMATICS FOR AI":                   {"credits": 3, "abs_diff": 4},
    "Object Oriented Analysis and Design":  {"credits": 3, "abs_diff": 2},
    "SOCIAL SCIENCES AND PROFESIONAL ETHICS":{"credits": 4, "abs_diff": 1},
    "BASICS OF MANAGEMENT":                 {"credits": 3, "abs_diff": 1},
    "Computer Networks":                    {"credits": 3, "abs_diff": 3},
    "Computer Organization and Architecture":{"credits": 4, "abs_diff": 3},
    "Database Information Systems":         {"credits": 3, "abs_diff": 2},
    "Machine Learning":                     {"credits": 3, "abs_diff": 3},
    "Theory of Computation":                {"credits": 4, "abs_diff": 4},
    "Compiler Design":                      {"credits": 3, "abs_diff": 4},
    "Cryptography":                         {"credits": 3, "abs_diff": 4},
    "Emerging Technologies for CS":         {"credits": 3, "abs_diff": 2},
    "Operating System":                     {"credits": 3, "abs_diff": 3},
    "Software Engineering (Minor)":         {"credits": 3, "abs_diff": 2},
    "Blockchain Technologies":              {"credits": 3, "abs_diff": 3},
    "Deep Learning":                        {"credits": 3, "abs_diff": 4},
    "Operating System (Minor)":             {"credits": 3, "abs_diff": 3},
    "Data Structures (Minor)":              {"credits": 3, "abs_diff": 3},
    "Artificial Intelligence (AI)":         {"credits": 3, "abs_diff": 3},
    "Computer and Network Security (CNS)":  {"credits": 3, "abs_diff": 3},
    "Digital Image Processing (DIP)":       {"credits": 3, "abs_diff": 3},
    "Parallel and Distributed Computing (PDC)":{"credits": 3, "abs_diff": 4},
    "Malware Analysis":                     {"credits": 3, "abs_diff": 4},
    "Wireless Networks (WN)":               {"credits": 3, "abs_diff": 2},
    "Data Structures":                      {"credits": 4, "abs_diff": 3},
    "Computer Vision":                      {"credits": 3, "abs_diff": 4},
    "Decentralized Learning":               {"credits": 3, "abs_diff": 4},
    "Full Stack Development":               {"credits": 3, "abs_diff": 2},
    "Information Retrieval":                {"credits": 3, "abs_diff": 3},
    "Natural Language Processing":          {"credits": 3, "abs_diff": 4},
    "Topics in Operating System":           {"credits": 3, "abs_diff": 4},
    # CHEM
    "Introduction to Chemical Engineering": {"credits": 3, "abs_diff": 1},
    "Chemical Engineering Thermodynamics-I":{"credits": 4, "abs_diff": 3},
    "Chemical Process Calculations":        {"credits": 4, "abs_diff": 3},
    "Process Instrumentation":              {"credits": 3, "abs_diff": 2},
    "Computer Aided Numerical Methods":     {"credits": 4, "abs_diff": 2},
    "Chemical Engineering Thermodynamics-II":{"credits": 4, "abs_diff": 4},
    "Conventional & Alternate Energy Resources":{"credits": 3, "abs_diff": 2},
    "Mechanical Operations":                {"credits": 4, "abs_diff": 2},
    "Fundamentals of Materials Science and Engineering":{"credits": 3, "abs_diff": 2},
    "Heat Transfer Operations":             {"credits": 4, "abs_diff": 3},
    "Industrial Pollution Abatement":       {"credits": 4, "abs_diff": 2},
    "Mass Transfer-I":                      {"credits": 4, "abs_diff": 3},
    "AI and ML in Chemical Engineering":    {"credits": 4, "abs_diff": 3},
    "Chemical Reaction Engineering-II":     {"credits": 4, "abs_diff": 4},
    "Mass Transfer-II":                     {"credits": 4, "abs_diff": 4},
    "Process Dynamics and Control":         {"credits": 4, "abs_diff": 4},
    "Process Safety and Hazards":           {"credits": 4, "abs_diff": 2},
    "DL":                                   {"credits": 3, "abs_diff": 4},
    "Big Data":                             {"credits": 3, "abs_diff": 3},
    "IoT":                                  {"credits": 3, "abs_diff": 2},
    "Advanced Process Control":             {"credits": 4, "abs_diff": 4},
    "CFD in Chemical Engineering":          {"credits": 4, "abs_diff": 4},
    "Mechanical Design of Process Equipment":{"credits": 4, "abs_diff": 3},
    "Nano-materials & Characterization":    {"credits": 3, "abs_diff": 3},
    "Optimization of Chemical Processes":   {"credits": 3, "abs_diff": 4},
    "Polymer Science and Technology":       {"credits": 3, "abs_diff": 2},
    "Process Integration":                  {"credits": 3, "abs_diff": 3},
    "Solid & Hazard Waste Management":      {"credits": 3, "abs_diff": 2},
    "Petroleum, Petrochemicals & Natural Gas Engineering":{"credits": 3, "abs_diff": 2},
    "Non-Conventional Energy Sources":      {"credits": 3, "abs_diff": 2},
    # AIDE
    "PROBLEM SOLVING WITH C":               {"credits": 2, "abs_diff": 2},
    "Discrete Structure":                   {"credits": 3, "abs_diff": 3},
    "Social Sciences and Professional Ethics":{"credits": 4, "abs_diff": 1},
    "Algorithm Design":                     {"credits": 3, "abs_diff": 4},
    "Automata Theory":                      {"credits": 4, "abs_diff": 4},
    "Digital Systems and Computer Architecture":{"credits": 4, "abs_diff": 3},
    "Foundations of Data Science":          {"credits": 3, "abs_diff": 2},
    "Artificial Neural Network":            {"credits": 3, "abs_diff": 3},
    "Data Communication and Networks":      {"credits": 3, "abs_diff": 2},
    "Database Management System":           {"credits": 3, "abs_diff": 2},
    "Introduction to Artificial Intelligence":{"credits": 3, "abs_diff": 2},
    "Introduction to Compiler Design":      {"credits": 3, "abs_diff": 4},
    "Big Data Analytics":                   {"credits": 3, "abs_diff": 3},
    "Data Mining and Warehousing":          {"credits": 3, "abs_diff": 2},
    "Image Processing":                     {"credits": 3, "abs_diff": 3},
    "Principles of Machine Learning":       {"credits": 3, "abs_diff": 3},
    "High Performance Computing":           {"credits": 3, "abs_diff": 4},
    "Information Security":                 {"credits": 3, "abs_diff": 3},
    # Sem 8
    "Internship": {"credits": 20, "abs_diff": 2},
    "Project":    {"credits": 6,  "abs_diff": 2},
    "Elective":   {"credits": 3,  "abs_diff": 2},
}

GRADE_TO_MARK = {"AA": 10, "AB": 9, "BB": 8, "BC": 7, "CC": 6, "CD": 5, "DD": 4, "FP": 0}
SEM8_NAMES    = {"Internship", "Project", "Elective"}
DEFAULT_INFO  = {"credits": 3, "abs_diff": 2}


def get_info(name):
    return SUBJECTS_DATA.get(name, DEFAULT_INFO)


def grade_mark_to_letter(gm):
    """Convert continuous grade mark (6-10) to letter grade."""
    gm = float(gm)
    if gm == 10 :  return "AA"
    if gm >= 9 :  return "AB"
    if gm >= 8 :  return "BB"
    if gm >= 7 :  return "BC"
    if gm >= 6 :  return "CC"
    if gm >= 5 :  return "CD"
    if gm >= 4 :  return "DD"
    return "FP"


def sgpa(subjects):
    tw = sum(s["grade_mark"] * s["credits"] for s in subjects)
    tc = sum(s["credits"] for s in subjects)
    return round(tw / tc, 4) if tc else 0.0


def cgpa(all_sgpas):
    return round(sum(all_sgpas) / len(all_sgpas), 4) if all_sgpas else 0.0


def trend_slope(sgpas):
    n = len(sgpas)
    if n < 2: return 0.0
    x  = list(range(1, n + 1))
    xm = sum(x) / n
    ym = sum(sgpas) / n
    num = sum((xi - xm) * (yi - ym) for xi, yi in zip(x, sgpas))
    den = sum((xi - xm) ** 2 for xi in x)
    return round(num / den, 4) if den else 0.0


def ml_predict(fv):
    """
    4-model ensemble (RF + GB + ET + Ridge).
    Returns grade_encoded (float 6-10), then caller rounds to letter.
    """
    arr = np.array(fv, dtype=float).reshape(1, -1)
    arr_imp = imputer.transform(arr)
    arr_sc  = scaler.transform(arr_imp)

    rf_pred    = rf_model.predict(arr_imp)[0]
    gb_pred    = gb_model.predict(arr_imp)[0]
    et_pred    = et_model.predict(arr_imp)[0]
    ridge_pred = ridge_model.predict(arr_sc)[0]
    pred = (W['rf'] * rf_pred + W['gb'] * gb_pred +
        W['et'] * et_pred + W['ridge'] * ridge_pred) / W_TOTAL


    intr_enc = fv[8]
    alignment = fv[11]

    if intr_enc <= 2:
        pred -= 0.5

    if alignment < 0:
        pred -= 0.4

    if intr_enc <= 2 and alignment < 0:
        pred -= 0.3


    return float(np.clip(pred, 6.0, 10.0))


@app.route('/')
def index():
    return render_template('index.html')


@app.route('/api/predict', methods=['POST'])
def predict():
    try:
        d = request.get_json(force=True)

        previous_sgpas = [float(s) for s in d['previous_sgpas']]
        beh            = d['behavioral']
        subjects_in    = d['subjects']

        # ── Global behavioural features ───────────────────────────
        jee   = float(beh.get('jee_percentile',  50.0))
        scr   = float(beh.get('screen_time',       4.0))
        std   = float(beh.get('study_hours',        2.0))
        exam  = float(beh.get('exam_study_hours',   3.0))
        seat  = int(beh.get('seating',           3))   # 1/3/5
        parti = int(beh.get('participation',     2))   # 1/2/4/5
        peer  = int(beh.get('peer_behaviour',    2))   # 1/2/4/5
        ca    = int(beh.get('career_alignment',  4))   # 1/2/4/5
        ec    = int(beh.get('extracurriculars',  2))   # 1/2/4/5

        # ── Trend + projected SGPA ────────────────────────────────
        hist   = previous_sgpas[:-1] if len(previous_sgpas) > 1 else previous_sgpas
        slope  = trend_slope(hist)
        p_sgpa = float(np.clip(previous_sgpas[-1] + slope, 0, 10)) \
                 if previous_sgpas else STATS['predicted_sgpa_mean']

        results = []
        for s in subjects_in:
            name    = s['name']
            is_sem8 = bool(s.get('is_sem8', name in SEM8_NAMES))
            info    = get_info(name)
            cred    = info['credits']
            abs_d   = info['abs_diff']

            # ── Sem 8: manual grade entry ─────────────────────────
            if is_sem8:
                mg = s.get('grade_manual', 'BB')
                results.append({
                    "name": name,
                    "grade": mg,
                    "grade_mark": float(GRADE_TO_MARK.get(mg, 8)),
                    "credits": cred,
                    "source": "manual"
                })
                continue

            # ── Regular subject ───────────────────────────────────
            filled = bool(s.get('filled', False))
            att    = int(s.get('attendance',  3))  if filled else 3   # 1-4
            intr   = int(s.get('interest',    4))  if filled else 4   # 1/2/4/5
            tch    = int(s.get('teacher',     3))  if filled else 3   # 1/3/4/6
            diff   = int(s.get('difficulty',  2))  if filled else 2   # 1/2/4/5

            # ── Derived features (match training data columns) ────
            # relative diff  = predicted_sgpa − abs_diff  (approximated)
            rel_diff   = p_sgpa - abs_d
            # relative allignment = career_alignment − difficulty  (approx)
            rel_align  = ca - diff
            # relative ec  = ec_encoded − abs_diff
            rel_ec     = ec - abs_d
            # interest and alignment w career encoded
            interest_x_alignment = intr * rel_align

            # Feature vector — ORDER must match FEATURE_COLS exactly:
            # JEE, screen_time, study_hrs, exam_hrs,
            # seat, parti, peer, att, intr, tch,
            # rel_diff, rel_align, rel_ec, predicted_sgpa
            fv = [jee, scr, std, exam,
                  seat, parti, peer,
                  att, intr, tch,
                  rel_diff, rel_align, rel_ec,
                  p_sgpa,  interest_x_alignment]

            gm = ml_predict(fv)
            results.append({
                "name":       name,
                "grade":      grade_mark_to_letter(gm),
                "grade_mark": round(gm, 2),
                "credits":    cred,
                "source":     "model" if filled else "default"
            })

        cur_sgpa  = sgpa([{"grade_mark": r["grade_mark"], "credits": r["credits"]}
                          for r in results])
        all_sgpas = previous_sgpas + [cur_sgpa]
        cur_cgpa  = cgpa(all_sgpas)

        return jsonify({
            "success":      True,
            "subjects":     results,
            "current_sgpa": round(cur_sgpa, 4),
            "cgpa":         round(cur_cgpa, 4)
        })

    except Exception as e:
        return jsonify({
            "success": False,
            "error":   str(e),
            "trace":   traceback.format_exc()
        }), 500


if __name__ == '__main__':
    import os

    if __name__ == "__main__":
       port = int(os.environ.get("PORT", 5000))
       app.run(host="0.0.0.0", port=port)