"""doc 단위 atomic 코퍼스 교체 (PR-DocSrv-Hierarchical-Decomposition-1 c5/c6).

legacy 윈도우 청크 → hier_section leaf 청크로 검색 코퍼스 교체(in_corpus 토글).
- 물리 삭제 없음(in_corpus 플래그만). 부분 ivfflat 이 자동 반영.
- G5 precondition(doc-local): hier leaf>0 + 모든 leaf embedding 보유(doc-local 100%) + parent 무결성(dangling 0).
- 단일 트랜잭션 atomic. 실패/precond 미충족 → 변경 0(legacy 유지).
- rollback: in_corpus 역토글(아래 rollback_doc_corpus).
"""
from __future__ import annotations
from sqlalchemy import text
from sqlalchemy.ext.asyncio import AsyncSession

CHUNKER_VERSION = "hier-rule-v1"


async def precheck(session: AsyncSession, doc_id: int) -> dict:
    row = (await session.execute(text("""
        SELECT
          count(*) FILTER (WHERE source_type='hier_section' AND is_leaf) AS hier_leaves,
          count(*) FILTER (WHERE source_type='hier_section' AND is_leaf AND embedding IS NOT NULL) AS hier_leaves_emb,
          count(*) FILTER (WHERE source_type='legacy' AND in_corpus) AS legacy_active,
          count(*) FILTER (WHERE source_type='hier_section' AND parent_id IS NOT NULL
              AND parent_id NOT IN (SELECT id FROM document_chunks WHERE doc_id=:d AND source_type='hier_section')) AS dangling
        FROM document_chunks WHERE doc_id=:d"""), {"d": doc_id})).one()
    leaves, leaves_emb = row.hier_leaves, row.hier_leaves_emb
    doc_local_100 = leaves > 0 and leaves_emb == leaves
    ok = doc_local_100 and row.dangling == 0
    return {
        "doc_id": doc_id, "hier_leaves": leaves, "hier_leaves_embedded": leaves_emb,
        "doc_local_embed_100": doc_local_100, "legacy_active": row.legacy_active,
        "dangling_parent": row.dangling, "precond_ok": ok,
        "reason": None if ok else (
            "no_hier_leaves" if leaves == 0 else
            "embed_incomplete" if not doc_local_100 else
            "dangling_parent"),
    }


async def replace_doc_corpus(session: AsyncSession, doc_id: int, *, dry_run: bool = True) -> dict:
    pc = await precheck(session, doc_id)
    pc["dry_run"] = dry_run
    if not pc["precond_ok"]:
        pc["action"] = "aborted"
        return pc
    if dry_run:
        pc["action"] = "dry_run"
        pc["would_deactivate_legacy"] = pc["legacy_active"]
        pc["would_activate_hier_leaves"] = pc["hier_leaves"]
        return pc
    # atomic 교체 (단일 트랜잭션)
    deact = (await session.execute(text(
        "UPDATE document_chunks SET in_corpus=false WHERE doc_id=:d AND source_type='legacy' AND in_corpus=true"),
        {"d": doc_id})).rowcount
    act = (await session.execute(text(
        "UPDATE document_chunks SET in_corpus=true WHERE doc_id=:d AND source_type='hier_section'"
        " AND chunker_version=:cv AND is_leaf=true AND embedding IS NOT NULL AND in_corpus=false"),
        {"d": doc_id, "cv": CHUNKER_VERSION})).rowcount
    await session.commit()
    pc.update({"action": "replaced", "legacy_deactivated": deact, "hier_activated": act})
    return pc


async def rollback_doc_corpus(session: AsyncSession, doc_id: int) -> dict:
    """교체 역토글 (legacy 복귀, hier 비활성)."""
    act = (await session.execute(text(
        "UPDATE document_chunks SET in_corpus=true WHERE doc_id=:d AND source_type='legacy' AND in_corpus=false"),
        {"d": doc_id})).rowcount
    deact = (await session.execute(text(
        "UPDATE document_chunks SET in_corpus=false WHERE doc_id=:d AND source_type='hier_section' AND in_corpus=true"),
        {"d": doc_id})).rowcount
    await session.commit()
    return {"doc_id": doc_id, "action": "rolled_back", "legacy_reactivated": act, "hier_deactivated": deact}