scripts/generate_curriculum.py

"""
scripts/generate_curriculum.py
Expands the seed curriculum with additional generated items and writes
data/T3.1_Math_Tutor/curriculum_full.json

Run once before starting the app:
    python3 scripts/generate_curriculum.py
"""
from __future__ import annotations
import json, random, sys
from pathlib import Path

SEED_PATH = Path("data/T3.1_Math_Tutor/curriculum_seed.json")
OUT_PATH  = Path("data/T3.1_Math_Tutor/curriculum_full.json")

# ── Generators ────────────────────────────────────────────────────────────────

def _counting_items(n_start: int) -> list[dict]:
    items = []
    for count in range(1, 11):
        diff = max(1, round(count / 2))
        for label in ["apple","star","ball","flower"]:
            items.append({
                "id":       f"cnt_gen_{count}_{label}",
                "skill":    "counting",
                "difficulty": diff,
                "answer_int": count,
                "stem_en":  f"How many {label}s do you see?",
                "stem_fr":  f"Combien de {label}s vois-tu ?",
                "stem_kin": f"Ubona {label} zingahe?",
                "stem_sw":  f"Unaona {label} ngapi?",
                "visual":   f"{label}_{count}",
            })
    return items


def _addition_items() -> list[dict]:
    items = []
    for a in range(0, 8):
        for b in range(0, 8):
            if a + b > 10:
                continue
            diff = max(1, min(9, a + b))
            items.append({
                "id":       f"add_gen_{a}_{b}",
                "skill":    "addition",
                "difficulty": diff,
                "answer_int": a + b,
                "stem_en":  f"{a} + {b} = ?",
                "stem_fr":  f"{a} + {b} = ?",
                "stem_kin": f"{a} + {b} = ?",
                "stem_sw":  f"{a} + {b} = ?",
                "visual":   "",
            })
    return items


def _subtraction_items() -> list[dict]:
    items = []
    for a in range(1, 11):
        for b in range(0, a + 1):
            diff = max(2, min(9, a))
            items.append({
                "id":       f"sub_gen_{a}_{b}",
                "skill":    "subtraction",
                "difficulty": diff,
                "answer_int": a - b,
                "stem_en":  f"{a} - {b} = ?",
                "stem_fr":  f"{a} - {b} = ?",
                "stem_kin": f"{a} - {b} = ?",
                "stem_sw":  f"{a} - {b} = ?",
                "visual":   "",
            })
    return items


def _number_sense_items() -> list[dict]:
    items = []
    for n in range(1, 10):
        items.append({
            "id": f"ns_gen_after_{n}",
            "skill": "number_sense", "difficulty": max(1, n//2),
            "answer_int": n + 1,
            "stem_en": f"What number comes after {n}?",
            "stem_fr": f"Quel nombre vient après {n} ?",
            "stem_kin": f"Ni umubare ukomeza {n}?",
            "stem_sw":  f"Nambari gani inakuja baada ya {n}?",
            "visual": "",
        })
        items.append({
            "id": f"ns_gen_before_{n+1}",
            "skill": "number_sense", "difficulty": max(1, n//2),
            "answer_int": n,
            "stem_en": f"What number comes before {n+1}?",
            "stem_fr": f"Quel nombre vient avant {n+1} ?",
            "stem_kin": f"Ni umubare uba imbere ya {n+1}?",
            "stem_sw":  f"Nambari gani inakuja kabla ya {n+1}?",
            "visual": "",
        })
    return items


# ── Main ──────────────────────────────────────────────────────────────────────

def main():
    with open(SEED_PATH, "r", encoding="utf-8") as f:
        seed = json.load(f)

    existing_ids = {item["id"] for item in seed}

    generated = (
        _counting_items(len(seed))
        + _addition_items()
        + _subtraction_items()
        + _number_sense_items()
    )

    # De-duplicate against seed
    new_items = [i for i in generated if i["id"] not in existing_ids]
    full      = seed + new_items
    random.shuffle(full)

    OUT_PATH.parent.mkdir(parents=True, exist_ok=True)
    with open(OUT_PATH, "w", encoding="utf-8") as f:
        json.dump(full, f, ensure_ascii=False, indent=2)

    print(f"✅ Generated {len(full)} items ({len(seed)} seed + {len(new_items)} new)")
    print(f"   Saved to {OUT_PATH}")


if __name__ == "__main__":
    main()