aps-agent/server/importers/sql_pack.py

# ============================================================
# MySQL/Navicat SQL 数据包解析（moduleId: importer-sql-pack, 可重生 ✅）
# 流式读取超大 dump（如松岳 mesdb_pro.sql ~1GB），抽取排产相关表 → flex*
# ============================================================
from __future__ import annotations

import os
import re
from datetime import datetime
from typing import Any, Iterable

# 表 → 角色（松岳 MES / 通用命名）
_TABLE_KIND: dict[str, str] = {
    "pl_order": "orders",
    "r_plan_order": "orders",
    "pl_order_bom": "bom",
    "md_material": "materials",
    "md_item": "materials",
    "md_equipment": "equipment",
    "em_devices": "equipment",
    "r_production_craftl": "routing",
    "c_project_process_route": "routing",
    "md_craftl": "craft",
    "r_production_procedure": "op_equip",
    "pl_work_order": "work_orders",
}

_CREATE_RE = re.compile(r"CREATE TABLE\s+`([^`]+)`", re.I)
_INSERT_RE = re.compile(r"INSERT INTO\s+`([^`]+)`\s+VALUES\s*", re.I)

# 单表预览/导入上限（防 1GB 全量撑爆内存）
_DEFAULT_LIMITS = {
    "orders": 800,
    "materials": 4000,
    "bom": 8000,
    "routing": 8000,          # 其它工艺上限
    "routing_prefer": 25000,  # 订单相关工艺路线优先配额
    "equipment": 800,
    "craft": 2000,
    "op_equip": 4000,
    "work_orders": 3000,
}


def _split_sql_values(blob: str) -> list[tuple[str, bool]]:
    """拆分 INSERT VALUES 元组内的字段 → (文本, 是否引号字符串)。"""
    out: list[tuple[str, bool]] = []
    cur: list[str] = []
    i = 0
    in_str = False
    quoted = False
    n = len(blob)
    while i < n:
        ch = blob[i]
        if in_str:
            if ch == "\\" and i + 1 < n:
                cur.append(blob[i + 1])
                i += 2
                continue
            if ch == "'":
                if i + 1 < n and blob[i + 1] == "'":
                    cur.append("'")
                    i += 2
                    continue
                in_str = False
                i += 1
                continue
            cur.append(ch)
            i += 1
            continue
        if ch == "'":
            in_str = True
            quoted = True
            i += 1
            continue
        if ch == ",":
            out.append(("".join(cur).strip(), quoted))
            cur = []
            quoted = False
            i += 1
            continue
        cur.append(ch)
        i += 1
    if cur or blob.endswith(","):
        out.append(("".join(cur).strip(), quoted))
    return out


def _parse_value(raw: str, *, quoted: bool = False) -> Any:
    s = (raw or "").strip()
    if not s or s.upper() == "NULL":
        return None
    if quoted:
        return s  # 引号字段保持字符串（料号 030129001 不能丢前导 0）
    if (s.startswith("'") and s.endswith("'")) or (s.startswith('"') and s.endswith('"')):
        return s[1:-1]
    try:
        if "." in s:
            return float(s)
        return int(s)
    except ValueError:
        return s


def _iter_insert_tuples(line: str) -> Iterable[list[Any]]:
    """从一行或多行 INSERT 语句产出行值列表。"""
    m = _INSERT_RE.search(line)
    if not m:
        return
    rest = line[m.end():].strip()
    if rest.endswith(";"):
        rest = rest[:-1]
    # 多个 (...),(...)
    i = 0
    n = len(rest)
    while i < n:
        while i < n and rest[i] in " \t\r\n,":
            i += 1
        if i >= n or rest[i] != "(":
            break
        depth = 0
        in_str = False
        j = i
        while j < n:
            ch = rest[j]
            if in_str:
                if ch == "\\" and j + 1 < n:
                    j += 2
                    continue
                if ch == "'":
                    if j + 1 < n and rest[j + 1] == "'":
                        j += 2
                        continue
                    in_str = False
                j += 1
                continue
            if ch == "'":
                in_str = True
                j += 1
                continue
            if ch == "(":
                depth += 1
            elif ch == ")":
                depth -= 1
                if depth == 0:
                    inner = rest[i + 1:j]
                    yield [_parse_value(x, quoted=q) for x, q in _split_sql_values(inner)]
                    i = j + 1
                    break
            j += 1
        else:
            break


def _parse_create_columns(block: str) -> list[str]:
    """解析 CREATE TABLE 列名；支持一行多列（测试/精简 dump）与 Navicat 一行一列。"""
    cols: list[str] = []
    for line in block.splitlines():
        s = line.strip().upper()
        if s.startswith(("PRIMARY ", "UNIQUE ", "KEY ", "INDEX ", "CONSTRAINT ")):
            continue
        if s.startswith(")"):
            break
        for m in re.finditer(r"`([^`]+)`\s+[a-zA-Z]", line):
            name = m.group(1)
            if name.upper() in ("PRIMARY", "UNIQUE", "KEY", "INDEX"):
                continue
            cols.append(name)
    return cols


def _due_str(v: Any) -> str:
    if v is None:
        return ""
    if isinstance(v, datetime):
        return v.strftime("%Y-%m-%d")
    s = str(v).strip()
    m = re.match(r"(\d{4}-\d{2}-\d{2})", s)
    return m.group(1) if m else s[:10]


def _row_dict(cols: list[str], vals: list[Any]) -> dict[str, Any]:
    d: dict[str, Any] = {}
    for i, c in enumerate(cols):
        if i < len(vals):
            d[c] = vals[i]
    return d


def scan_sql_pack(
    path: str,
    *,
    limits: dict[str, int] | None = None,
) -> dict[str, Any]:
    """流式扫描 SQL dump，返回按 kind 分组的行 + 表统计。"""
    lim = {**_DEFAULT_LIMITS, **(limits or {})}
    schemas: dict[str, list[str]] = {}
    rows_by_table: dict[str, list[dict[str, Any]]] = {}
    counts: dict[str, int] = {}
    creating: str | None = None
    create_buf: list[str] = []

    size = os.path.getsize(path) if os.path.isfile(path) else 0
    try:
        from server.agent_core.progress import emit_thinking
    except Exception:
        def emit_thinking(*_a, **_k):  # type: ignore
            return None

    emit_thinking("读取 SQL 数据包", f"{os.path.basename(path)} · {size / (1024 ** 2):.0f} MB", pct=1)
    lines_read = 0
    bytes_read = 0
    last_emit = 0
    # 订单上的工艺路线编码：dump 里 pl_order 早于 r_production_craftl，可优先抽取
    prefer_craftl: set[str] = set()
    routing_pref: list[dict[str, Any]] = []
    routing_other: list[dict[str, Any]] = []
    routing_table = "r_production_craftl"
    with open(path, "r", encoding="utf-8", errors="replace") as f:
        for line in f:
            lines_read += 1
            bytes_read += len(line.encode("utf-8", errors="ignore"))
            if creating is not None:
                create_buf.append(line)
                if "ENGINE" in line.upper() and line.strip().endswith(";"):
                    schemas[creating] = _parse_create_columns("".join(create_buf))
                    creating = None
                    create_buf = []
                continue

            cm = _CREATE_RE.search(line)
            if cm:
                name = cm.group(1)
                if name in _TABLE_KIND:
                    creating = name
                    create_buf = [line]
                    emit_thinking("解析表结构", name, pct=min(40, 5 + len(schemas)))
                continue

            if "INSERT INTO" not in line.upper():
                if lines_read - last_emit >= 400_000:
                    last_emit = lines_read
                    pct = min(85.0, 10 + (bytes_read / max(size, 1)) * 70)
                    emit_thinking("扫描 SQL 全文", f"已读 {lines_read // 10000 / 10:.1f} 万行", pct=pct)
                continue
            im = _INSERT_RE.search(line)
            if not im:
                continue
            table = im.group(1)
            kind = _TABLE_KIND.get(table)
            if not kind:
                continue
            cols = schemas.get(table) or []
            if not cols:
                continue
            # 支持 INSERT 跨行：VALUES 在下一行，直到分号结束
            insert_blob = line
            while ";" not in insert_blob:
                nxt = f.readline()
                if not nxt:
                    break
                lines_read += 1
                bytes_read += len(nxt.encode("utf-8", errors="ignore"))
                insert_blob += nxt
            before = counts.get(table, 0)
            for vals in _iter_insert_tuples(insert_blob):
                counts[table] = counts.get(table, 0) + 1
                row = _row_dict(cols, vals)
                if kind == "orders":
                    cc = str(row.get("craftl_code") or "").strip()
                    if cc:
                        prefer_craftl.add(cc)
                    bucket = rows_by_table.setdefault(table, [])
                    if len(bucket) < lim.get("orders", 800):
                        bucket.append(row)
                elif kind == "routing":
                    cc = str(row.get("craftl_code") or row.get("process_id") or "").strip()
                    if cc and cc in prefer_craftl:
                        if len(routing_pref) < lim.get("routing_prefer", 25000):
                            routing_pref.append(row)
                    elif len(routing_other) < lim.get("routing", 8000):
                        routing_other.append(row)
                else:
                    bucket = rows_by_table.setdefault(table, [])
                    cap = lim.get(kind, 1000)
                    if len(bucket) < cap:
                        bucket.append(row)
            if counts.get(table, 0) and before == 0:
                emit_thinking("抽取业务表", f"{table} → {kind}", pct=min(90, 20 + len(counts) * 5))
            if lines_read - last_emit >= 400_000:
                last_emit = lines_read
                pct = min(90.0, 10 + (bytes_read / max(size, 1)) * 75)
                emit_thinking(
                    "扫描 SQL 全文",
                    f"已读 {lines_read // 10000 / 10:.1f} 万行 · 订单工艺 {len(prefer_craftl)} 条",
                    pct=pct,
                )

    if routing_pref or routing_other:
        rows_by_table[routing_table] = routing_pref + routing_other
        emit_thinking(
            "工艺优先装载",
            f"订单相关 {len(routing_pref)} 步 · 其它 {len(routing_other)} 步 · 覆盖工艺编码 {len(prefer_craftl)}",
            status="done", pct=91,
        )

    emit_thinking("SQL 扫描完成", f"命中 {len(counts)} 张相关表", status="done", pct=92)
    return {
        "path": path,
        "size": size,
        "schemas": {k: v for k, v in schemas.items()},
        "counts": counts,
        "rowsByTable": rows_by_table,
        "preferCraftl": sorted(prefer_craftl),
        "kinds": sorted({_TABLE_KIND[t] for t in counts}),
    }


def _map_orders(rows: list[dict]) -> list[dict]:
    out = []
    for r in rows:
        if str(r.get("is_delete") or "0") not in ("0", "0.0", "", "None"):
            continue
        ono = str(r.get("code") or r.get("order_code") or "").strip()
        pc = str(r.get("material_code") or r.get("product_code") or "").strip()
        if not ono or not pc:
            continue
        qty = float(r.get("quantity") or r.get("quatity") or r.get("all_quantity") or 0)
        if qty <= 0:
            qty = 1
        due = _due_str(r.get("planned_end_time") or r.get("end_time") or r.get("actual_end_time"))
        if not due:
            due = datetime.now().strftime("%Y-%m-%d")
        out.append({
            "orderNo": ono,
            "productCode": pc,
            "productName": str(r.get("material_name") or r.get("product_name") or pc),
            "quantity": int(qty) if qty == int(qty) else qty,
            "dueDate": due,
            "priority": int(float(r.get("level") or 5) or 5),
            "customerName": str(r.get("customer_business_code") or r.get("dept") or "松岳现场"),
            "status": "RELEASED",
            "drawingNo": str(r.get("drawing_code") or r.get("figure_no") or ""),
            "machineModel": str(r.get("machine_model") or r.get("model") or ""),
            "craftlCode": str(r.get("craftl_code") or "").strip(),
        })
    return out


def _map_materials(rows: list[dict]) -> list[dict]:
    out = []
    seen: set[str] = set()
    for r in rows:
        if str(r.get("is_delete") or "0") not in ("0", "0.0", "", "None"):
            continue
        code = str(r.get("code") or "").strip()
        name = str(r.get("name") or code).strip()
        if not code or code in seen:
            continue
        seen.add(code)
        raw_type = str(r.get("type") or r.get("classify_name") or "").upper()
        type_cn = str(r.get("type") or "")
        if "毛坯" in name or "原料" in type_cn or raw_type in ("RM", "RAW"):
            mtype = "RAW_MATERIAL"
        elif "半" in type_cn or raw_type in ("SF", "SEMI"):
            mtype = "SEMI_FINISHED"
        elif "成品" in type_cn or raw_type in ("FG", "FINISHED"):
            mtype = "FINISHED_PRODUCT"
        else:
            # 松岳料号习惯：01* 毛坯/原料，其余先按原料，有工艺时再升为半成品
            mtype = "RAW_MATERIAL" if code.startswith("01") else "SEMI_FINISHED"
        sourcing = "BUY" if mtype == "RAW_MATERIAL" or "毛坯" in name else "MAKE"
        out.append({
            "code": code, "name": name, "type": mtype,
            "sourcingType": sourcing,
            "unit": str(r.get("unit") or "件").strip('"') or "件",
            "stock": float(r.get("stock") or 0),
            "inTransit": 0, "safetyStock": 0, "procurementLeadTime": 0,
            "spec": str(r.get("specification_type") or r.get("figure_no") or ""),
        })
    return out


def _map_equipment(rows: list[dict]) -> list[dict]:
    out = []
    seen: set[str] = set()
    for r in rows:
        if str(r.get("is_delete") or r.get("delete_flag") or "0") not in ("0", "0.0", "", "None"):
            continue
        code = str(r.get("code") or r.get("devices_code") or "").strip()
        name = str(r.get("name") or r.get("devices_name") or code).strip()
        if not code or code in seen:
            continue
        seen.add(code)
        out.append({
            "code": code, "name": name,
            "capabilities": ["GENERAL"],
            "opStdTime": {"GENERAL": 1.0},
            "movable": False, "moveTimeMin": 0,
            "zone": "ZONE-A", "adaptableMolds": [],
            "availabilityRate": 0.95, "status": "RUNNING",
        })
    return out


def _map_routing(rows: list[dict], craft_product: dict[str, str]) -> list[dict]:
    """工艺步骤挂到成品：优先 craftl_code→订单/工艺主数据料号，禁止用中间件料号冒充成品。"""
    out = []
    for r in rows:
        if str(r.get("is_delete") or r.get("del_flag") or "0") not in ("0", "0.0", "", "None"):
            continue
        op = str(r.get("procedure_code") or r.get("process_code") or "").strip()
        if not op:
            continue
        craft = str(r.get("craftl_code") or r.get("process_id") or "").strip()
        out_mat = str(r.get("output_material_code") or r.get("product_code") or "").strip()
        # 成品优先：工艺编码映射 > 输出料号（仅当映射缺失）
        pc = craft_product.get(craft) or out_mat or craft
        if not pc:
            continue
        try:
            seq = int(float(r.get("sort_no") or r.get("process_sort") or r.get("serial_number") or 10))
        except (TypeError, ValueError):
            seq = 10
        if seq <= 0:
            seq = 10
        std = r.get("working_hours") or r.get("process_shift_output")
        try:
            std_f = float(std) if std not in (None, "") else None
        except (TypeError, ValueError):
            std_f = None
        # 工时若像小时则转分钟
        if std_f is not None and std_f > 0 and std_f < 24:
            std_f = round(std_f * 60, 2)
        if std_f is None or std_f <= 0:
            std_f = 1.0
            src = "推断"
        else:
            src = "实测"
        pname = str(r.get("craftl_name") or r.get("output_material_name") or pc)
        op_name = str(r.get("procedure_name") or r.get("process_name") or op)
        from server.aps_domain.sourcing import infer_op_sourcing
        sourcing = infer_op_sourcing(op, op_name)
        out.append({
            "productCode": pc,
            "productName": pname,
            "seq": seq,
            "operationCode": op,
            "operationName": op_name,
            "requireMold": False,
            "stdTimePerUnit": std_f,
            "stdTimeSource": src,
            "craftlCode": craft,
            "outputMaterialCode": out_mat,
            "sourcingType": sourcing,
            "isExternal": sourcing == "OUTSOURCE",
        })
    return out


def _expand_routing_for_orders(
    orders: list[dict],
    steps: list[dict],
) -> list[dict]:
    """同一 craftl 下的步骤，按订单成品料号再挂一份（多订单共用工艺）。"""
    by_craft: dict[str, list[dict]] = {}
    for s in steps:
        cc = str(s.get("craftlCode") or "").strip()
        if cc:
            by_craft.setdefault(cc, []).append(s)
    out: list[dict] = []
    seen: set[tuple] = set()
    for o in orders:
        pc = str(o.get("productCode") or "").strip()
        cc = str(o.get("craftlCode") or "").strip()
        if not pc:
            continue
        src = by_craft.get(cc) or [s for s in steps if s.get("productCode") == pc]
        for s in src:
            key = (pc, s.get("seq"), s.get("operationCode"))
            if key in seen:
                continue
            seen.add(key)
            row = dict(s)
            row["productCode"] = pc
            row["productName"] = o.get("productName") or pc
            out.append(row)
    # 保留未能挂到订单的步骤（其它产品）
    for s in steps:
        key = (s.get("productCode"), s.get("seq"), s.get("operationCode"))
        if key not in seen:
            out.append(s)
            seen.add(key)
    return out


def _map_bom(rows: list[dict], order_product: dict[str, str]) -> list[dict]:
    out = []
    for r in rows:
        if str(r.get("is_delete") or "0") not in ("0", "0.0", "", "None"):
            continue
        ono = str(r.get("order_code") or "").strip()
        mc = str(r.get("material_code") or "").strip()
        pc = order_product.get(ono) or str(r.get("machine_model") or "").strip()
        qty = float(r.get("composition_quantity") or r.get("consumption_quantity") or r.get("quantity") or 0)
        if not pc or not mc or qty <= 0:
            continue
        out.append({
            "productCode": pc, "materialCode": mc, "quantity": qty,
            "consumeOp": str(r.get("procedure_code") or ""),
            "isKey": qty >= 1,
            "materialName": str(r.get("material_name") or mc),
        })
    return out


def _apply_op_equip(equip: list[dict], rows: list[dict]) -> None:
    """用工序-设备关系补全 capabilities。"""
    by_code = {e["code"]: e for e in equip}
    for r in rows:
        if str(r.get("is_delete") or "0") not in ("0", "0.0", "", "None"):
            continue
        ec = str(r.get("equipment_code") or "").strip()
        op = str(r.get("procedure_code") or "").strip()
        if not ec or not op or ec not in by_code:
            continue
        e = by_code[ec]
        caps = list(e.get("capabilities") or [])
        if "GENERAL" in caps and len(caps) == 1:
            caps = []
        if op not in caps:
            caps.append(op)
        e["capabilities"] = caps or ["GENERAL"]
        std = dict(e.get("opStdTime") or {})
        if op not in std:
            std[op] = 1.0
        if "GENERAL" in std and op != "GENERAL" and len(std) > 1:
            std.pop("GENERAL", None)
        e["opStdTime"] = std


def sql_pack_to_flex(scan: dict[str, Any]) -> dict[str, Any]:
    """把 scan 结果映射为可写入 world 的 flex* 批次。"""
    rbt = scan.get("rowsByTable") or {}
    craft_product: dict[str, str] = {}
    for r in rbt.get("md_craftl") or []:
        code = str(r.get("code") or "").strip()
        pc = str(r.get("product_code") or "").strip()
        if code and pc:
            craft_product[code] = pc

    orders = _map_orders((rbt.get("pl_order") or []) + (rbt.get("r_plan_order") or []))
    # 订单上的 craftl→成品料号优先（现场口径，覆盖 md_craftl 偏差）
    for o in orders:
        cc = str(o.get("craftlCode") or "").strip()
        pc = str(o.get("productCode") or "").strip()
        if cc and pc:
            craft_product[cc] = pc

    order_product = {o["orderNo"]: o["productCode"] for o in orders}
    # 订单产品补进物料
    materials = _map_materials((rbt.get("md_material") or []) + (rbt.get("md_item") or []))
    mat_codes = {m["code"] for m in materials}
    for o in orders:
        if o["productCode"] not in mat_codes:
            materials.append({
                "code": o["productCode"], "name": o.get("productName") or o["productCode"],
                "type": "FINISHED_PRODUCT", "unit": "件",
                "stock": 0, "inTransit": 0, "safetyStock": 0, "procurementLeadTime": 0, "spec": "",
            })
            mat_codes.add(o["productCode"])

    equipment = _map_equipment((rbt.get("md_equipment") or []) + (rbt.get("em_devices") or []))
    _apply_op_equip(equipment, rbt.get("r_production_procedure") or [])
    routing_raw = _map_routing(
        (rbt.get("r_production_craftl") or []) + (rbt.get("c_project_process_route") or []),
        craft_product,
    )
    routing = _expand_routing_for_orders(orders, routing_raw)
    # 设备能力：把订单工艺工序挂到设备（MES 设备表常无工序能力列）
    ops_needed = sorted({str(s.get("operationCode") or "") for s in routing if s.get("operationCode")})
    if ops_needed and equipment:
        for e in equipment:
            caps = [c for c in (e.get("capabilities") or []) if c and c != "GENERAL"]
            if not caps:
                # MES 设备台账常无工序能力列：按已挂工艺放开，否则齐备度永远卡 NO_CAPABLE
                e["capabilities"] = list(ops_needed)
                e["opStdTime"] = {op: float((e.get("opStdTime") or {}).get(op) or 1.0) for op in ops_needed}
    bom = _map_bom(rbt.get("pl_order_bom") or [], order_product)

    pcs_ord = {o["productCode"] for o in orders}
    pcs_rt = {s["productCode"] for s in routing}
    return {
        "orders": orders,
        "materials": materials,
        "equipment": equipment,
        "routing": routing,
        "bom": bom,
        "stats": {
            "orders": len(orders), "materials": len(materials),
            "equipment": len(equipment), "routing": len(routing), "bom": len(bom),
            "orderProducts": len(pcs_ord),
            "routedProducts": len(pcs_ord & pcs_rt),
            "missingRouting": len(pcs_ord - pcs_rt),
            "rawCounts": scan.get("counts") or {},
        },
    }


def sync_flex_to_classic_master(world: dict[str, Any]) -> dict[str, int]:
    """把 flex* 投影到主数据页读的经典表（materials/routings/…），避免工艺模型仍显示演示厂。"""
    summary: dict[str, int] = {}

    def _next(table: str) -> int:
        items = world.get(table) or []
        return max((x.get("id", 0) for x in items if isinstance(x.get("id"), int)), default=0) + 1

    # 物料
    mats = world.setdefault("materials", [])
    code_to_id: dict[str, int] = {str(m.get("code")): int(m["id"]) for m in mats if m.get("id") is not None}
    for row in world.get("flexMaterials") or []:
        code = str(row.get("code") or "").strip()
        if not code:
            continue
        payload = {
            "code": code,
            "name": row.get("name") or code,
            "spec": row.get("spec") or "",
            "type": row.get("type") or "FINISHED_PRODUCT",
            "unit": row.get("unit") or "件",
            "productFamily": row.get("productFamily") or "",
            "safetyStock": float(row.get("safetyStock") or 0),
            "procurementLeadTime": float(row.get("procurementLeadTime") or 0),
            "stock": float(row.get("stock") or 0),
            "inTransit": float(row.get("inTransit") or 0),
            "expectedArrivalDate": row.get("expectedArrivalDate") or "",
            "sourcingType": row.get("sourcingType") or "",
            "sourcingTypeSource": row.get("sourcingTypeSource") or ("IMPORT" if row.get("sourcingType") else ""),
            "status": "ACTIVE",
        }
        if code in code_to_id:
            ex = next(m for m in mats if m.get("code") == code)
            ex.update(payload)
        else:
            mid = _next("materials")
            mats.append({"id": mid, **payload})
            code_to_id[code] = mid
        summary["materials"] = summary.get("materials", 0) + 1

    # 工序库
    ops = world.setdefault("operations", [])
    op_code_to_id: dict[str, int] = {str(o.get("code")): int(o["id"]) for o in ops if o.get("id") is not None}
    for row in world.get("flexOperations") or []:
        code = str(row.get("code") or "").strip()
        if not code:
            continue
        if code in op_code_to_id:
            continue
        oid = _next("operations")
        from server.aps_domain.sourcing import infer_op_sourcing
        sourcing = infer_op_sourcing(code, row.get("name"))
        ops.append({
            "id": oid, "code": code, "name": row.get("name") or code,
            "type": "EXTERNAL" if sourcing == "OUTSOURCE" else "INTERNAL",
            "sourcingType": sourcing,
            "standardTime": float((row.get("changeoverMin") or 0) or 0) / 60.0 or 0.5,
        })
        op_code_to_id[code] = oid
        summary["operations"] = summary.get("operations", 0) + 1

    # 工艺路线 + 步骤（按产品）
    routings = world.setdefault("routings", [])
    steps = world.setdefault("routingSteps", [])
    live_pcs = {str(r.get("productCode") or "") for r in (world.get("flexRoutings") or [])}
    if live_pcs:
        drop_pids = {code_to_id[pc] for pc in live_pcs if pc in code_to_id}
        drop_rids = {r["id"] for r in routings if r.get("productId") in drop_pids}
        if drop_rids:
            world["routings"] = [r for r in routings if r["id"] not in drop_rids]
            world["routingSteps"] = [s for s in steps if s.get("routingId") not in drop_rids]
            routings = world["routings"]
            steps = world["routingSteps"]

    by_pc: dict[str, list[dict]] = {}
    for row in world.get("flexRoutings") or []:
        pc = str(row.get("productCode") or "").strip()
        if pc:
            by_pc.setdefault(pc, []).append(row)
    for pc, rows in by_pc.items():
        pid = code_to_id.get(pc)
        if not pid:
            continue
        rid = _next("routings")
        routings.append({
            "id": rid, "productId": pid, "version": "V1",
            "versionName": f"{pc}工艺", "isDefault": True, "status": "ACTIVE",
        })
        rows_sorted = sorted(rows, key=lambda r: int(r.get("seq") or 0))
        prev = None
        for i, row in enumerate(rows_sorted):
            op = str(row.get("operationCode") or "").strip()
            if op and op not in op_code_to_id:
                oid = _next("operations")
                ops.append({
                    "id": oid, "code": op, "name": row.get("operationName") or op,
                    "type": "EXTERNAL" if (
                        row.get("isExternal") or
                        str(row.get("sourcingType") or "").upper() == "OUTSOURCE" or
                        str(op).upper().startswith("WZ")
                    ) else "INTERNAL",
                    "sourcingType": row.get("sourcingType") or (
                        "OUTSOURCE" if str(op).upper().startswith("WZ") else "MAKE"
                    ),
                    "standardTime": float(row.get("stdTimePerUnit") or 1) / 60.0,
                })
                op_code_to_id[op] = oid
            oid = op_code_to_id.get(op)
            if not oid:
                continue
            sid = _next("routingSteps")
            std = float(row.get("stdTimePerUnit") or 1.0)
            from server.aps_domain.sourcing import infer_op_sourcing
            sourcing = row.get("sourcingType") or infer_op_sourcing(op, row.get("operationName"))
            is_ext = bool(row.get("isExternal")) or sourcing == "OUTSOURCE"
            # 同步工序主数据 type
            for o in ops:
                if o.get("id") == oid:
                    o["type"] = "EXTERNAL" if is_ext else o.get("type") or "INTERNAL"
                    o["sourcingType"] = sourcing
                    break
            steps.append({
                "id": sid, "routingId": rid, "operationId": oid,
                "sequenceNo": i + 1, "prevStepId": prev,
                "setupTime": 10, "runTimePerUnit": std, "waitTime": 0,
                "transferTime": 0, "isExternal": is_ext,
                "sourcingType": sourcing,
            })
            prev = sid
            summary["routingSteps"] = summary.get("routingSteps", 0) + 1
            if is_ext:
                summary["externalSteps"] = summary.get("externalSteps", 0) + 1
        summary["routings"] = summary.get("routings", 0) + 1

    # BOM
    boms = world.setdefault("boms", [])
    bom_items = world.setdefault("bomItems", [])
    bom_by_pc: dict[str, list[dict]] = {}
    for row in world.get("flexBom") or []:
        pc = str(row.get("productCode") or "").strip()
        if pc:
            bom_by_pc.setdefault(pc, []).append(row)
    for pc, rows in bom_by_pc.items():
        pid = code_to_id.get(pc)
        if not pid:
            continue
        # 清旧
        old = [b["id"] for b in boms if b.get("productId") == pid]
        if old:
            world["boms"] = [b for b in boms if b["id"] not in old]
            world["bomItems"] = [bi for bi in bom_items if bi.get("bomId") not in old]
            boms = world["boms"]
            bom_items = world["bomItems"]
        bid = _next("boms")
        boms.append({
            "id": bid, "productId": pid, "version": "V1",
            "versionName": f"{pc}BOM", "isDefault": True, "status": "ACTIVE",
        })
        for row in rows:
            mc = str(row.get("materialCode") or "").strip()
            if not mc:
                continue
            if mc not in code_to_id:
                mid = _next("materials")
                mats.append({
                    "id": mid, "code": mc, "name": row.get("materialName") or mc,
                    "spec": "", "type": "RAW_MATERIAL", "unit": row.get("unit") or "件",
                    "productFamily": "", "safetyStock": 0, "procurementLeadTime": 0,
                    "stock": 0, "inTransit": 0, "status": "ACTIVE",
                })
                code_to_id[mc] = mid
            bom_items.append({
                "id": _next("bomItems"), "bomId": bid,
                "materialId": code_to_id[mc],
                "quantity": float(row.get("quantity") or 1),
                "operationId": None, "isKeyMaterial": False,
            })
            summary["bomItems"] = summary.get("bomItems", 0) + 1
        summary["boms"] = summary.get("boms", 0) + 1

    # 资源树：无工厂时挂一个「现场」占位，设备挂到柔性资源页即可
    if (world.get("flexEquipment") or []) and not (world.get("factories") or []):
        world["factories"] = [{
            "id": 1, "code": "SITE", "name": "现场工厂",
            "timezone": "Asia/Shanghai", "address": "", "status": "ACTIVE",
        }]
        world["workshops"] = [{
            "id": 1, "factoryId": 1, "code": "WS1", "name": "生产车间", "status": "ACTIVE",
        }]
        world["lines"] = [{
            "id": 1, "workshopId": 1, "code": "FLEX", "name": "柔性能力池",
            "capacityPerDay": 9999, "taktTime": 1, "efficiencyFactor": 1.0,
            "status": "ACTIVE", "alternativeLineIds": [],
        }]
        summary["factories"] = 1

    params = world.setdefault("flexParams", {})
    params["demoDataCleared"] = True
    params["siteProfile"] = params.get("siteProfile") or "sql-pack"
    return summary


def apply_sql_pack_to_world(world: dict[str, Any], flex: dict[str, Any], *, replace: bool = True) -> dict[str, int]:
    """将 SQL 映射结果写入 world flex*，并投影到主数据页经典表。"""
    from server.state.seed import ensure_flex_seed, purge_demo_residue
    # 先清演示残留，再保证键存在（不会回填假数据）
    purge_demo_residue(world)
    world.setdefault("flexParams", {})["demoDataCleared"] = True
    ensure_flex_seed(world)
    summary: dict[str, int] = {}

    def _next(table: str) -> int:
        items = world.get(table) or []
        return max((x.get("id", 0) for x in items if isinstance(x.get("id"), int)), default=0) + 1

    if replace:
        for k in ("flexOrders", "flexMaterials", "flexEquipment", "flexRoutings", "flexBom"):
            world[k] = []
        # 清掉演示厂班组，否则现场工序（NZ*/WZ*）会被 NO_TEAM 卡死
        world["flexTeams"] = []
        # 经典主数据也整表换新，禁止智能控制器与现场料号并存
        for k in (
            "materials", "boms", "bomItems", "operations", "routings", "routingSteps",
            "lineProducts", "workstationOperations", "changeoverMatrix",
            "factories", "workshops", "lines", "workstations", "equipment",
        ):
            world[k] = []

    # materials first
    fm = world.setdefault("flexMaterials", [])
    for row in flex.get("materials") or []:
        ex = next((m for m in fm if m.get("code") == row["code"]), None)
        if ex:
            ex.update(row)
        else:
            fm.append({"id": _next("flexMaterials"), **row})
        summary["materials"] = summary.get("materials", 0) + 1

    fo = world.setdefault("flexOrders", [])
    for row in flex.get("orders") or []:
        ex = next((o for o in fo if o.get("orderNo") == row["orderNo"]), None)
        if ex:
            ex.update(row)
        else:
            fo.append({"id": _next("flexOrders"), **row})
        summary["orders"] = summary.get("orders", 0) + 1

    fe = world.setdefault("flexEquipment", [])
    for row in flex.get("equipment") or []:
        ex = next((e for e in fe if e.get("code") == row["code"]), None)
        if ex:
            ex.update(row)
        else:
            fe.append({"id": _next("flexEquipment"), **row})
        summary["equipment"] = summary.get("equipment", 0) + 1

    fr = world.setdefault("flexRoutings", [])
    for row in flex.get("routing") or []:
        fr[:] = [s for s in fr if not (
            s.get("productCode") == row["productCode"] and s.get("seq") == row["seq"]
            and s.get("operationCode") == row["operationCode"]
        )]
        fr.append(dict(row))
        summary["routing"] = summary.get("routing", 0) + 1

    fb = world.setdefault("flexBom", [])
    for row in flex.get("bom") or []:
        fb[:] = [s for s in fb if not (
            s.get("productCode") == row["productCode"]
            and s.get("materialCode") == row["materialCode"]
        )]
        fb.append(dict(row))
        summary["bom"] = summary.get("bom", 0) + 1

    # 默认班次
    if not world.get("flexCalendar"):
        world["flexCalendar"] = [{
            "shiftCode": "D", "startTime": "08:00", "endTime": "17:00",
            "breaks": [{"start": "12:00", "end": "13:00"}],
            "workdays": [1, 2, 3, 4, 5],
        }]
        summary["calendar"] = 1

    # 工序库
    ops = world.setdefault("flexOperations", [])
    seen_op = {o.get("code") for o in ops}
    all_ops: list[str] = []
    for row in flex.get("routing") or []:
        code = row.get("operationCode")
        if code and code not in seen_op:
            ops.append({"code": code, "name": row.get("operationName") or code,
                        "isBottleneck": False, "changeoverMin": 10})
            seen_op.add(code)
            summary["operations"] = summary.get("operations", 0) + 1
        if code and code not in all_ops:
            all_ops.append(code)

    # 现场综合班组：覆盖 SQL 工艺全部工序，避免演示厂班组资格卡死
    if all_ops:
        teams = world.setdefault("flexTeams", [])
        plant = next((t for t in teams if t.get("code") == "T-PLANT"), None)
        if plant is None:
            plant = {
                "code": "T-PLANT", "name": "现场综合班组",
                "memberCount": max(8, min(40, len(world.get("flexEquipment") or []) or 8)),
                "supportOps": [], "skillLevel": "L3",
            }
            teams.append(plant)
            summary["teams"] = 1
        for code in all_ops:
            if code not in plant["supportOps"]:
                plant["supportOps"].append(code)

    # 投影到主数据「工艺模型」页（经典 materials/routings）
    classic = sync_flex_to_classic_master(world)
    for k, v in classic.items():
        summary[f"classic_{k}"] = v

    # 投影到订单管理页（salesOrders）
    try:
        from server.aps_domain.orders import sync_flex_orders_to_sales
        n = sync_flex_orders_to_sales(world)
        if n:
            summary["salesOrders"] = n
    except Exception:
        pass

    # 工序/物料供应类型标注（WZ*=委外，毛坯=采购）
    try:
        from server.aps_domain.sourcing import annotate_world_sourcing
        summary["sourcing"] = annotate_world_sourcing(world)
    except Exception:
        pass

    return summary


def preview_sql_file(path: str) -> dict[str, Any]:
    """给 folder_pack / 分析用的轻量预览。"""
    scan = scan_sql_pack(path)
    flex = sql_pack_to_flex(scan)
    st = flex["stats"]
    return {
        "filename": os.path.basename(path),
        "kind": "sql-pack",
        "kindCn": "SQL 数据包",
        "size": scan["size"],
        "okCount": st["orders"] + st["materials"] + st["routing"] + st["equipment"],
        "errorCount": 0,
        "tables": [
            {"table": t, "kind": _TABLE_KIND.get(t, ""), "rawCount": c,
             "loaded": len((scan.get("rowsByTable") or {}).get(t) or [])}
            for t, c in sorted((scan.get("counts") or {}).items())
        ],
        "fieldMapText": (
            f"订单{st['orders']} · 物料{st['materials']} · 工艺{st['routing']} · "
            f"设备{st['equipment']} · BOM{st['bom']}"
        ),
        "samples": [
            f"订单 {o['orderNo']}：{o['productCode']}×{o['quantity']} 交期{o['dueDate']}"
            for o in (flex.get("orders") or [])[:3]
        ],
        "flex": flex,
        "canSchedule": st["orders"] > 0 and (st["routing"] > 0 or st["equipment"] > 0),
    }