#!/usr/bin/env python3
"""Deterministic classifier for the frozen 8-item probe seed.

Lane: sol / Triage (everyday Q&A and classification).

Rule (evidence-keyword driven, stdlib only):
  Category:
    - archive_fix  -> evidence describes a defect / omission in deployed content:
                      hero mark, funding table, leaderboard restat, export rows,
                      readme env vars.
    - app_pr       -> evidence is a publish / submit action.
    - general      -> everything else (batch / handover / coordination).
  Decision:
    - approve      -> proposed_action applies a single corrective change
                      (update / add / fix / publish) and names no second owner.
    - table        -> proposed_action batches a group or hands a file to a named
                      second owner (batch ... digest / hand ... to Timmy).

Input:  wizard-work/sol/seed-2026-10-29.json  (frozen, 8 pending items)
Output: wizard-work/sol/Triage-2026-11-01.json (manifest)
"""
import json
import re
from pathlib import Path

REPO = Path(__file__).resolve().parent.parent.parent
SEED = REPO / "wizard-work" / "sol" / "seed-2026-10-29.json"
OUT = REPO / "wizard-work" / "sol" / "Triage-2026-11-01.json"

# Evidence keywords -> category. Checked in order; first match wins.
CATEGORY_RULES = [
    ("archive_fix", ["hero", "funding table", "restats", "export drops",
                     "readme omits", "readme", "funding table rows lack"]),
    ("app_pr", ["publish", "submit", "release-notes"]),
]


def classify_category(evidence: str) -> str:
    low = evidence.lower()
    for cat, kws in CATEGORY_RULES:
        if any(k in low for k in kws):
            return cat
    return "general"


def classify_decision(proposed_action: str) -> str:
    low = proposed_action.lower()
    # Handover / batch items name a second owner or a group -> table.
    if re.search(r"\b(batch|hand|record|send)\b", low):
        return "table"
    return "approve"


def main() -> None:
    seed = json.loads(SEED.read_text(encoding="utf-8"))
    pending = seed["pending"]

    items = []
    for s in pending:
        evidence = s["evidence"]
        summary = s["summary"]
        proposed_action = s["proposed_action"]
        items.append({
            "id": s["id"],
            "summary": summary,
            "category": classify_category(evidence),
            "decision": classify_decision(proposed_action),
            "rationale": (
                f"evidence='{evidence}' -> category rule matched '{classify_category(evidence)}';"
                f" proposed_action='{proposed_action}' -> decision rule matched '{classify_decision(proposed_action)}'."
            ),
            "evidence": evidence,
            "source": "seed-2026-10-29.json (frozen)",
        })

    decision_counts = {}
    categories = {}
    for it in items:
        decision_counts[it["decision"]] = decision_counts.get(it["decision"], 0) + 1
        categories[it["category"]] = categories.get(it["category"], 0) + 1

    manifest = {
        "artifact": "Triage-2026-11-01.json",
        "source": "wizard-work/sol/seed-2026-10-29.json (frozen 8-item probe set)",
        "generated_at": "2026-11-01T00:00:00Z",
        "lane": "Triage - everyday Q&A and classification",
        "pending_in_seed": len(pending),
        "summary": "Triage of 8 pending items in the frozen seed, classified by "
                   "evidence-keyword rules.",
        "decision_counts": decision_counts,
        "categories": categories,
        "items": items,
    }
    OUT.write_text(json.dumps(manifest, indent=2) + "\n", encoding="utf-8")
    print(f"Wrote {OUT}")
    print(f"decision_counts={decision_counts} categories={categories}")


if __name__ == "__main__":
    main()