"""Build data/ergo-knowledge-base.json from the raw datasource dumps.

Run:  uv run python scripts/build_knowledge_base.py
"""

from __future__ import annotations

import json
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]
PARAMS_FILE = ROOT / "docs" / "datasources" / "common-assessment-parameters.json"
ISSUES_FILE = ROOT / "docs" / "datasources" / "issue-based-outcomes.json"
OUT_FILE = ROOT / "data" / "ergo-knowledge-base.json"


def _none_if_empty(value):
    if value is None:
        return None
    if isinstance(value, str) and value.strip() == "":
        return None
    return value


def main() -> None:
    raw_params = json.loads(PARAMS_FILE.read_text())
    raw_issues = json.loads(ISSUES_FILE.read_text())

    issues_by_outcome = {row["outcome"]: row for row in raw_issues}

    parameters = []
    for p in raw_params:
        options = []
        for o in p["options"]:
            key = o["option"]
            risk_title = _none_if_empty(o.get("riskTitle"))
            overall_risk = _none_if_empty(o.get("overallRisk"))
            is_good_habit = risk_title is None and overall_risk is None

            issue = issues_by_outcome.get(key)
            if issue is None:
                raise SystemExit(
                    f"Missing issue-based-outcomes entry for option key: {key}"
                )
            posture_score = issue["postureScore"]
            # Parameter file is the authoritative visual rubric — if it says
            # good habit, force risk_level to null. Otherwise take the issue
            # file's risk_level.
            if is_good_habit:
                risk_level = None
            else:
                risk_level = _none_if_empty(issue.get("riskLevel"))
                if risk_level is None:
                    raise SystemExit(
                        f"{key}: risky in parameter file but issue riskLevel is null"
                    )

            options.append(
                {
                    "key": key,
                    "label": o["optionText"],
                    "reference_image": o["optionImage"],
                    "is_good_habit": is_good_habit,
                    "risk_level": risk_level,
                    "posture_score": posture_score,
                    "risk_title": risk_title,
                    "overall_risk": overall_risk,
                }
            )

        selection_mode = "multi" if p.get("has_multi_select") else "single"
        parameters.append(
            {
                "id": p["id"],
                "group": p["group"],
                "order": p["order"],
                "selection_mode": selection_mode,
                "parameter_text": p["parameterText"],
                "question_text": p["question_text"],
                "options": options,
            }
        )

    payload = {"version": "1.0", "parameters": parameters}

    OUT_FILE.parent.mkdir(parents=True, exist_ok=True)
    OUT_FILE.write_text(json.dumps(payload, indent=2) + "\n")
    print(f"Wrote {OUT_FILE} ({len(parameters)} parameters, "
          f"{sum(len(p['options']) for p in parameters)} options)")


if __name__ == "__main__":
    main()