src/config.py

"""
config.py — Centralised configuration for the Smart MCQ Solver.

All constants, label mappings, model paths, and
example questions live here so they are easy to change in one place.
"""

import os
import torch

# ── Device ────────────────────────────────────────────────────────────────────
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"

# ── Label mapping (must match training) ───────────────────────────────────────
LABEL2ID = {"A": 0, "B": 1, "C": 2, "D": 3, "E": 4}
ID2LABEL = {v: k for k, v in LABEL2ID.items()}
NUM_LABELS = 5

# ── Tokenisation ──────────────────────────────────────────────────────────────
MAX_LEN = 384

# ── Model source ──────────────────────────────────────────────────────────────
# Set MODEL_SOURCE env var to "local" to load from disk instead of the Hub.
# Defaults to "hub" because the Space repo does not ship model weights —
# they live in the separate Pro-aryan00/roberta-mcq-ft model repo.
MODEL_SOURCE = os.environ.get("MODEL_SOURCE", "hub").lower()

# HuggingFace Hub config (used when MODEL_SOURCE == "hub")
HF_USERNAME  = os.environ.get("HF_USERNAME", "Pro-aryan00")
HF_TOKEN     = os.environ.get("HF_TOKEN", None)
ROBERTA_HUB_REPO = f"{HF_USERNAME}/roberta-mcq-ft"

# Local model paths (used when MODEL_SOURCE == "local")
# Resolve relative to project root, not cwd
_PROJECT_ROOT = os.path.abspath(os.path.join(os.path.dirname(__file__), ".."))
ROBERTA_LOCAL_PATH = os.path.join(_PROJECT_ROOT, "models", "roberta-mcq-ft")

# Base checkpoint on HuggingFace Hub (always downloaded from Hub)
ROBERTA_BASE_CHECKPOINT = "roberta-base"

# ── UI colours for probability bars ───────────────────────────────────────────
BAR_COLORS = ["#a78bfa", "#60a5fa", "#34d399", "#f472b6", "#fb923c"]

# ── Example questions for the demo ────────────────────────────────────────────
# These are real test samples that RoBERTa answered correctly with 99%+ confidence.
EXAMPLES = [
    {
        # Sample 10 — SI base unit of time → B (correct: the second)
        "prompt": "Which of the following is correct? What is the SI base unit of time and how is it defined? carefully.",
        "A": "The SI base unit of time is the week, which is defined by measuring the electronic transition frequency of caesium atoms.",
        "B": "The SI base unit of time is the second, which is defined by measuring the electronic transition frequency of caesium atoms.",
        "C": "The SI base unit of time is the hour, which is defined by measuring the electronic transition frequency of caesium atoms.",
        "D": "The SI base unit of time is the day, which is defined by measuring the electronic transition frequency of caesium atoms.",
        "E": "The SI base unit of time is the minute, which is defined by measuring the electronic transition frequency of caesium atoms.",
    },
    {
        # Sample 9 — Maxwell's Demon → C (correct description)
        "prompt": "What is the Maxwell's Demon thought experiment?",
        "A": "A thought experiment in which a demon guards a microscopic trapdoor in a wall separating two parts of a container filled with different gases at equal temperatures. The demon selectively allows molecules to pass from one side to the other, causing an increase in temperature in one part and a decrease in temperature in the other, contrary to the second law of thermodynamics.",
        "B": "A thought experiment in which a demon guards a macroscopic trapdoor in a wall separating two parts of a container filled with different gases at different temperatures. The demon selectively allows molecules to pass from one side to the other, causing a decrease in temperature in one part and an increase in temperature in the other, in accordance with the second law of thermodynamics.",
        "C": "A thought experiment in which a demon guards a microscopic trapdoor in a wall separating two parts of a container filled with the same gas at equal temperatures. The demon selectively allows faster-than-average molecules to pass from one side to the other, causing a decrease in temperature in one part and an increase in temperature in the other, contrary to the second law of thermodynamics.",
        "D": "A thought experiment in which a demon guards a macroscopic trapdoor in a wall separating two parts of a container filled with the same gas at equal temperatures. The demon selectively allows faster-than-average molecules to pass from one side to the other, causing an increase in temperature in one part and a decrease in temperature in the other, contrary to the second law of thermodynamics.",
        "E": "A thought experiment in which a demon guards a microscopic trapdoor in a wall separating two parts of a container filled with the same gas at different temperatures. The demon selectively allows slower-than-average molecules to pass from one side to the other, causing a decrease in temperature in one part and an increase in temperature in the other, in accordance with the second law of thermodynamics.",
    },
    {
        # Sample 5 — Landau-Lifshitz-Gilbert equation → C (correct description)
        "prompt": "What is the Landau-Lifshitz-Gilbert equation used for in physics?",
        "A": "The Landau-Lifshitz-Gilbert equation is a differential equation used to describe the precessional motion of magnetization M in a liquid, and is commonly used in micromagnetics to model the effects of a magnetic field on ferromagnetic materials.",
        "B": "The Landau-Lifshitz-Gilbert equation is a differential equation used to describe the precessional motion of magnetization M in a solid, and is commonly used in astrophysics to model the effects of a magnetic field on celestial bodies.",
        "C": "The Landau-Lifshitz-Gilbert equation is a differential equation used to describe the precessional motion of magnetization M in a solid, and is commonly used in micromagnetics to model the effects of a magnetic field on ferromagnetic materials.",
        "D": "The Landau-Lifshitz-Gilbert equation is a differential equation used to describe the precessional motion of magnetization M in a solid, and is commonly used in macro-magnetics to model the effects of a magnetic field on ferromagnetic materials.",
        "E": "The Landau-Lifshitz-Gilbert equation is a differential equation used to describe the precessional motion of magnetization M in a liquid, and is commonly used in macro-magnetics to model the effects of a magnetic field on ferromagnetic materials.",
    },
]