aps-agent/server/aps_domain/importers.py

# ============================================================
# Excel/CSV 导入管线（moduleId: domain-importers, 可重生 ✅）
# 对齐 docs/product/demand-data-intake.md · features MD-04
# preview（P1 校验）→ commit（P2 确认卡落库）
# ============================================================
from __future__ import annotations

import csv
import io
import re
from typing import Any

World = dict[str, Any]

_HEADER_ALIASES: dict[str, str] = {
    "编码": "code", "code": "code", "设备编码": "code", "模具编码": "code",
    "设备编号": "code", "工装模具编号": "code", "模具编号": "code",
    "物料编码": "code", "物料代码": "code", "物料编号": "code",
    "ERP品号": "code", "erp品号": "code", "财务编号": "code",
    "固定资产编号": "code", "厂内编号": "code",
    "物品编码": "code", "存货编码": "code", "零件编码": "code",
    "产品编码": "productCode", "productcode": "productCode",
    "成品编码": "productCode", "料号": "productCode", "产品料号": "productCode",
    "父件编码": "productCode", "父项编码": "productCode", "父项产品编码": "productCode", "母件编码": "productCode",
    "名称": "name", "name": "name", "设备名称": "name", "物料简称": "name",
    "模具名称": "name", "工装模具名称": "name", "物料名称": "name",
    "产品名称": "productName", "物料描述": "productName", "productname": "productName",
    "类型": "type", "type": "type", "物料类型": "type", "物料分类": "type", "物料组名称": "type",
    "单位": "unit", "unit": "unit", "计量单位": "unit",
    "库存": "stock", "stock": "stock", "当前库存": "stock", "现有库存": "stock",
    "在途": "inTransit", "intransit": "inTransit", "安全库存": "safetyStock",
    "前置期": "procurementLeadTime", "采购前置期": "procurementLeadTime",
    "预计到货日期": "expectedArrivalDate", "预计到货": "expectedArrivalDate", "expectedarrivaldate": "expectedArrivalDate",
    "采购方式": "sourcingType", "供应类型": "sourcingType", "sourcingtype": "sourcingType",
    "客户": "customerName", "customername": "customerName", "客户名称": "customerName",
    "项目名称": "customerName", "项目": "customerName",
    "数量": "quantity", "quantity": "quantity", "单套用量": "quantity",
    "订单数量": "quantity", "计划订单数量": "quantity", "单位用量": "quantity", "用量": "quantity",
    "交期": "deliveryDate", "duedate": "deliveryDate", "交货期": "deliveryDate",
    "交货日期": "deliveryDate", "计划交期": "deliveryDate", "需求日期": "deliveryDate",
    "计划结束时间": "deliveryDate", "计划结束": "deliveryDate", "完成日期": "deliveryDate",
    "订单号": "orderNo", "orderno": "orderNo", "订单代码": "orderNo",
    "生产订单": "orderNo", "工单编号": "orderNo", "工单号": "orderNo",
    "优先级": "priority", "priority": "priority", "加急": "isRush", "isrush": "isRush",
    "等级": "customerLevel", "客户等级": "customerLevel",
    "工序编码": "operationCode", "operationcode": "operationCode",
    "工序编号": "operationCode", "工序": "operationCode",
    "工序名称": "name", "operationname": "name",
    "序号": "seq", "seq": "seq", "顺序": "seq", "排序号": "seq",
    "单件工时": "stdTimePerUnit", "stdtimeperunit": "stdTimePerUnit",
    "工时": "stdTimePerUnit", "标准工时": "stdTimePerUnit", "标准工时/分钟": "stdTimePerUnit",
    "需模具": "requireMold", "requiremold": "requireMold",
    "是否瓶颈": "isBottleneck", "isbottleneck": "isBottleneck", "瓶颈": "isBottleneck",
    "换型": "changeoverMin", "换型时间": "changeoverMin", "changeovermin": "changeoverMin",
    "能力": "capabilities", "capabilities": "capabilities", "可执行工序": "capabilities",
    "单件工时映射": "opStdTime", "opstdtime": "opStdTime",
    "可移动": "movable", "movable": "movable", "移动耗时": "moveTimeMin",
    "movetimemin": "moveTimeMin", "区域": "zone", "zone": "zone",
    "适配模具": "adaptableMolds", "adaptablemolds": "adaptableMolds",
    "适配设备": "adaptableEquipment", "adaptableequipment": "adaptableEquipment",
    "可动率": "availabilityRate", "availabilityrate": "availabilityRate",
    "状态": "status", "status": "status", "寿命上限": "lifeTotal", "lifetotal": "lifeTotal",
    "已用寿命": "lifeUsed", "lifeused": "lifeUsed",
    "子件编码": "materialCode", "子项编码": "materialCode", "子项物料编码": "materialCode", "组件编码": "materialCode",
    "materialcode": "materialCode",
    "关键料": "isKey", "iskey": "isKey", "是否关键件": "isKey",
    "消耗工序": "consumeOp", "consumeop": "consumeOp",
    "mes工序编号": "consumeOp", "MES工序编号": "consumeOp",
    "规格型号": "spec",
    "班次编码": "shiftCode", "shiftcode": "shiftCode", "班次名称": "name",
    "开始时间": "startTime", "starttime": "startTime", "结束时间": "endTime", "endtime": "endTime",
    "工作日": "workdays", "workdays": "workdays", "休息时段": "breaks", "breaks": "breaks",
    "启用": "enabled", "enabled": "enabled",
}

# 各类型「排产关键字段」：源表头别名 → 标准字段（用于动态识别展示）
_KIND_KEY_FIELDS: dict[str, tuple[tuple[str, ...], ...]] = {
    "orders": (
        ("productCode", "料号", "产品编码", "成品编码"),
        ("quantity", "订单数量", "计划订单数量", "数量"),
        ("deliveryDate", "计划结束时间", "完成日期", "交期"),
        ("customerName", "项目名称", "客户", "客户名称"),
        ("orderNo", "订单代码", "订单号", "生产订单"),
    ),
    "calendar": (
        ("shiftCode", "班次编码"),
        ("startTime", "开始时间"),
        ("endTime", "结束时间"),
    ),
    "materials": (
        ("code", "物料代码", "物料编码", "物料编号", "编码"),
        ("name", "物料名称", "名称"),
    ),
    "routing": (
        ("operationCode", "工序编号", "工序编码", "工序"),
        ("seq", "排序号", "序号", "顺序"),
        ("stdTimePerUnit", "标准工时/分钟", "标准工时", "工时", "单件工时"),
        ("productCode", "产品编码", "料号", "成品编码"),
    ),
    "equipment": (
        ("code", "设备编号", "设备编码", "编码"),
        ("name", "设备名称", "名称"),
        ("capabilities", "可执行工序", "能力"),
    ),
    "molds": (
        ("code", "工装模具编号", "模具编号", "模具编码", "编码"),
        ("name", "工装模具名称", "模具名称", "名称"),
        ("operationCode", "工序编号", "适用工序", "工序"),
    ),
    "bom": (
        ("materialCode", "物料编号", "物料编码", "子件编码"),
        ("quantity", "单位用量", "用量", "数量"),
        ("productCode", "成品编码", "产品编码", "料号"),
        ("orderNo", "工单编号", "订单代码"),
    ),
    "operations": (
        ("code", "工序编号", "工序编码", "编码"),
        ("name", "工序名称", "名称"),
    ),
}

_KIND_HINTS: dict[str, tuple[str, ...]] = {
    "orders": ("订单", "order", "销售订单"),
    "calendar": ("班历", "班次", "calendar", "shift"),
    "materials": ("物料", "material", "库存"),
    "bom": ("bom", "物料清单", "产品bom"),
    "routing": ("工艺路线", "routing", "路线"),
    "equipment": ("设备", "equipment", "机器"),
    "molds": ("模具", "mold", "工装"),
    "operations": ("工序", "operation", "工序库"),
    "zones": ("区域", "zone", "布局"),
}

IMPORT_KINDS = tuple(_KIND_HINTS.keys())


def _norm_header(h: str) -> str:
    raw = str(h or "").strip()
    key = re.sub(r"[\s\u3000]+", "", raw.lower())
    direct = _HEADER_ALIASES.get(raw) or _HEADER_ALIASES.get(key)
    if direct:
        return direct

    # 现场模板常在列名后追加「(必填)」「[分钟]」或星号；先去掉说明再匹配。
    cleaned = re.sub(r"[（(【\[].*?[）)】\]]", "", key)
    cleaned = re.sub(r"[＊*#：:_\-]+", "", cleaned)
    cleaned = re.sub(r"(?:必填|字段|列名)$", "", cleaned)
    return _HEADER_ALIASES.get(cleaned) or cleaned


def _truthy(v: Any) -> bool:
    if isinstance(v, bool):
        return v
    s = str(v or "").strip().lower()
    return s in ("1", "true", "yes", "y", "是", "有")


def _split_multi(v: Any) -> list[str]:
    if v is None or v == "":
        return []
    if isinstance(v, (list, tuple)):
        return [str(x).strip() for x in v if str(x).strip()]
    return [p.strip() for p in re.split(r"[,，;/|、]", str(v)) if p.strip()]


def _parse_op_std(v: Any) -> dict[str, float]:
    if isinstance(v, dict):
        return {str(k): float(val) for k, val in v.items()}
    out: dict[str, float] = {}
    for part in _split_multi(v):
        if ":" in part:
            k, t = part.split(":", 1)
            try:
                out[k.strip()] = float(t)
            except ValueError:
                continue
    return out


def detect_kind(name: str, headers: list[str]) -> str | None:
    blob = (name or "").lower()
    for kind, hints in _KIND_HINTS.items():
        if any(h.lower() in blob for h in hints):
            return kind
    hs = {h.lower() for h in headers}
    hdrs = {str(h or "").strip() for h in headers}
    mapped = {_norm_header(h) for h in headers if h}
    if (
        {"customername", "客户", "交期", "deliverydate", "订单代码", "料号", "订单数量"} & (hs | hdrs)
        or ({"productcode", "quantity"} <= mapped and ("deliverydate" in mapped or "customername" in mapped))
        or ("orderno" in mapped and "productcode" in mapped)
    ):
        return "orders"
    if {"starttime", "endtime"} <= mapped or {"开始时间", "结束时间"} <= hdrs:
        return "calendar"
    if "capabilities" in mapped or "可执行工序" in hdrs or "movable" in mapped or "设备编号" in hdrs:
        return "equipment"
    if "lifetotal" in mapped or "寿命上限" in hdrs or "工装模具编号" in hdrs:
        return "molds"
    if "isbottleneck" in mapped or "是否瓶颈" in hdrs:
        return "operations"
    if "operationcode" in mapped and ("seq" in mapped or "stdtimeperunit" in mapped):
        return "routing"
    if ("materialcode" in mapped or "物料编号" in hdrs) and ("quantity" in mapped or "单位用量" in hdrs):
        return "bom"
    if "stock" in mapped or "库存" in hdrs or "物料代码" in hdrs or "物料编码" in hdrs:
        return "materials"
    if "zone" in mapped and len(headers) <= 4:
        return "zones"
    return None


def detect_field_map(kind: str, headers_raw: list[str]) -> list[dict[str, str]]:
    """根据原始表头动态识别关键字段映射：源列 → 标准字段。"""
    specs = _KIND_KEY_FIELDS.get(kind) or ()
    found: list[dict[str, str]] = []
    seen_targets: set[str] = set()
    raw_list = [str(h or "").strip() for h in headers_raw if h]
    raw_set = set(raw_list)
    for group in specs:
        target = group[0]
        if target in seen_targets:
            continue
        hit = None
        # 1) 优先精确命中中文/英文别名（跳过 target 自身，避免 序号 抢在 排序号 前）
        for alias in group[1:]:
            if alias in raw_set:
                hit = alias
                break
        # 2) 再按规范化列名匹配
        if not hit:
            for h in raw_list:
                if _norm_header(h) == target:
                    hit = h
                    break
        if hit:
            seen_targets.add(target)
            found.append({"source": hit, "target": target})
    return found


def _sheet_as_code(sheet: str | None) -> str:
    s = str(sheet or "").strip()
    if not s or s.lower().startswith("sheet") or s in ("订单导出", "物料", "Equipment"):
        return ""
    return s


def _rows_from_csv(raw: bytes) -> tuple[list[str], list[dict[str, Any]]]:
    text = raw.decode("utf-8-sig", errors="replace")
    reader = csv.DictReader(io.StringIO(text))
    headers = [_norm_header(h) for h in (reader.fieldnames or [])]
    rows = []
    for r in reader:
        item = {_norm_header(k): (v.strip() if isinstance(v, str) else v) for k, v in r.items() if k}
        if any(str(v or "").strip() for v in item.values()):
            rows.append(item)
    return headers, rows


_XLSX_SOURCE_KEY = "__source__"
_XLSX_DIAGNOSTIC_KEY = "__diagnostic__"
_EMBEDDED_EQUIPMENT_SECTIONS = {"备品备件管理", "设备部件管理", "设备类型"}
_RAW_SUMMARY_PRIORITY = (
    "财务编号", "固定资产编号", "厂内编号", "设备名称",
    "设备型号", "出厂编号", "可执行工序", "工序名称",
)


def _is_blank(value: Any) -> bool:
    return value is None or (isinstance(value, str) and not value.strip())


def _jsonish_value(value: Any) -> Any:
    if value is None or isinstance(value, (bool, int, float)):
        return value
    if hasattr(value, "isoformat"):
        try:
            return value.isoformat()
        except (TypeError, ValueError):
            pass
    text = str(value).strip()
    return text if len(text) <= 120 else f"{text[:117]}..."


def _raw_row_summary(headers_raw: list[str], line: tuple[Any, ...], limit: int = 8) -> dict[str, Any]:
    """返回有限、可 JSON 序列化的原始行摘要；关键编码字段即使为空也保留。"""
    by_header: dict[str, Any] = {}
    for idx, header in enumerate(headers_raw):
        if not header or idx >= len(line):
            continue
        value = line[idx]
        if header not in by_header or (not _is_blank(value) and _is_blank(by_header[header])):
            by_header[header] = value

    result: dict[str, Any] = {}
    for header in _RAW_SUMMARY_PRIORITY:
        if header in by_header and len(result) < limit:
            result[header] = _jsonish_value(by_header[header])
    for header, value in by_header.items():
        if len(result) >= limit:
            break
        if header in result or _is_blank(value):
            continue
        result[header] = _jsonish_value(value)
    return result


def _set_mapped_value(item: dict[str, Any], key: str, value: Any) -> bool:
    """写入规范字段；返回 True 表示空的重复别名被安全忽略。"""
    if key not in item:
        item[key] = value
        return False
    if _is_blank(value):
        return not _is_blank(item[key])
    item[key] = value
    return False


def _diagnostic_counts(diagnostics: list[dict[str, Any]]) -> dict[str, int]:
    counts = {"ignored": 0, "warning": 0, "blocking": 0}
    for item in diagnostics:
        severity = str(item.get("severity") or "")
        if severity in counts:
            counts[severity] += 1
    return counts


def _row_source(sheet: str, excel_row: int, headers_raw: list[str], line: tuple[Any, ...],
                section: str = "") -> dict[str, Any]:
    return {
        "sheet": sheet,
        "excelRow": excel_row,
        "rawSummary": _raw_row_summary(headers_raw, line),
        "section": section,
    }

def _rows_from_xlsx(raw: bytes) -> list[dict[str, Any]]:
    from openpyxl import load_workbook

    wb = load_workbook(io.BytesIO(raw), read_only=True, data_only=True)
    sheets = []
    known_targets = set(_HEADER_ALIASES.values())
    for ws in wb.worksheets:
        data = list(ws.iter_rows(values_only=True))
        if not data:
            continue
        # 现场模板常在首行放标题/合并单元格（如 MOM 收集表的「物料清单」、
        # 分块「物料组 + 物料主数据」）：扫描规范化后命中已知字段映射最多
        # 的行作为表头（序号/编码/名称/单位/数量…），跳过标题行；
        # 平铺表第 1 行即表头时行为不变；无任何命中退回第 1 行。
        best_idx, best_score = 0, -1
        for idx, row in enumerate(data):
            if not row:
                continue
            score = sum(
                1 for cell in row
                if cell is not None and str(cell).strip() and _norm_header(cell) in known_targets
            )
            if score > best_score:
                best_idx, best_score = idx, score

        headers_raw = [str(cell).strip() if cell is not None else "" for cell in data[best_idx]]
        headers = [_norm_header(header) for header in headers_raw]
        flow_idx = next(
            (idx for idx, header in enumerate(headers_raw)
             if header and ("制造流程" in header or "工艺代码" in header or "工艺流程" in header)),
            None,
        )

        active_section = ""
        for prior in data[:best_idx + 1]:
            if prior and not _is_blank(prior[0]):
                active_section = str(prior[0]).strip()

        rows: list[dict[str, Any]] = []
        for excel_row, line in enumerate(data[best_idx + 1:], start=best_idx + 2):
            if not line or all(_is_blank(cell) for cell in line):
                continue
            if not _is_blank(line[0]):
                active_section = str(line[0]).strip()

            source = _row_source(ws.title, excel_row, headers_raw, line, active_section)
            item: dict[str, Any] = {}
            normalizations: list[dict[str, Any]] = []
            for idx, header in enumerate(headers):
                if not header or idx >= len(line):
                    continue
                value = line[idx]
                if isinstance(value, str):
                    value = value.strip()
                ignored_empty_alias = _set_mapped_value(item, header, value)
                if ignored_empty_alias and header == "code":
                    normalizations.append({
                        "code": "DUPLICATE_ALIAS_EMPTY_IGNORED",
                        "message": f"空的重复编码列「{headers_raw[idx]}」未覆盖已有编码",
                    })
            if normalizations:
                source["normalizations"] = normalizations

            sequence_values = []
            if flow_idx is not None:
                sequence_values = [
                    str(line[idx]).strip()
                    for idx in range(flow_idx, len(line))
                    if idx < len(line) and not _is_blank(line[idx])
                ]
            is_secondary_header = (
                flow_idx is not None
                and excel_row == best_idx + 2
                and len(sequence_values) >= 2
                and all(re.fullmatch(r"\d{1,2}", value) for value in sequence_values)
                and _is_blank(item.get("code"))
                and _is_blank(item.get("name"))
            )
            if is_secondary_header:
                source["rawSummary"] = {
                    **source.get("rawSummary", {}),
                    "制造流程序号": ",".join(sequence_values),
                }
                rows.append({
                    _XLSX_SOURCE_KEY: source,
                    _XLSX_DIAGNOSTIC_KEY: {
                        "kind": "routing",
                        "code": "SECONDARY_SEQUENCE_HEADER",
                        "severity": "ignored",
                        "message": "制造流程二级序号表头，已忽略",
                    },
                })
                continue

            is_equipment_sheet = "设备" in ws.title or "equipment" in ws.title.lower()
            if is_equipment_sheet and active_section in _EMBEDDED_EQUIPMENT_SECTIONS:
                rows.append({
                    _XLSX_SOURCE_KEY: source,
                    _XLSX_DIAGNOSTIC_KEY: {
                        "kind": "equipment",
                        "code": "EMBEDDED_SECTION_ROW",
                        "severity": "ignored",
                        "message": f"{active_section}为嵌入式非设备实例数据，已忽略",
                    },
                })
                continue

            if any(not _is_blank(value) for value in item.values()):
                item[_XLSX_SOURCE_KEY] = source
                rows.append(item)

        # 制造流程列（锐扬 MOM 02-工艺模型）：表头行后次行为 01..12 序号，
        # 数据行该列起每格 = 工艺代码（如 ZP01组装 / LM01POP拉铆）→ 合成
        # routing 视图，供 preview_file 按 routing 批校验/入库。
        if flow_idx is not None:
            flow_rows: list[dict[str, Any]] = []
            seq_map: dict[int, int] = {}
            sub = data[best_idx + 1] if best_idx + 1 < len(data) else None
            if sub is not None:
                for idx in range(flow_idx, len(sub)):
                    value = sub[idx]
                    if value is not None and str(value).strip().isdigit():
                        seq_map[idx] = int(str(value).strip())
            for excel_row, line in enumerate(data[best_idx + 2:], start=best_idx + 3):
                if not line or all(_is_blank(cell) for cell in line):
                    continue
                item: dict[str, Any] = {}
                for idx, header in enumerate(headers):
                    if not header or idx >= len(line):
                        continue
                    value = line[idx]
                    if isinstance(value, str):
                        value = value.strip()
                    _set_mapped_value(item, header, value)
                product_code = str(item.get("code") or item.get("productCode") or "").strip()
                if not product_code:
                    continue
                source = _row_source(ws.title, excel_row, headers_raw, line, active_section)
                for idx in range(flow_idx, len(headers)):
                    value = line[idx] if idx < len(line) else None
                    if _is_blank(value):
                        continue
                    raw_op = str(value).strip()
                    match = re.match(r"^([A-Za-z]{1,12}\d{0,6}[A-Za-z]{0,8})\s*(.*)$", raw_op)
                    operation_code = match.group(1) if match else raw_op
                    operation_name = (
                        match.group(2) if match and match.group(2) else raw_op
                    ).strip() or operation_code
                    sequence = seq_map.get(idx) or (idx - flow_idx + 1)
                    flow_rows.append({
                        "productCode": product_code,
                        "seq": sequence,
                        "operationCode": operation_code,
                        "operationName": operation_name,
                        _XLSX_SOURCE_KEY: source,
                    })
            if flow_rows:
                sheets.append({
                    "sheet": f"{ws.title}#routing",
                    "physicalSheet": ws.title,
                    "kind": "routing",
                    "headers": ["productCode", "seq", "operationCode", "operationName"],
                    "headersRaw": ["产品编码", "序号", "工序编码", "工序名称"],
                    "rows": flow_rows,
                })
        sheets.append({
            "sheet": ws.title,
            "physicalSheet": ws.title,
            "headers": headers,
            "headersRaw": headers_raw,
            "rows": rows,
        })
    return sheets


def _due_str(v: Any) -> str:
    if v is None:
        return ""
    if hasattr(v, "strftime"):
        return v.strftime("%Y-%m-%d")
    return str(v).strip()[:10]


def validate_batch(
    kind: str,
    rows: list[dict[str, Any]],
    world: World,
    *,
    soft: bool = False,
    sheet: str | None = None,
    product_by_order: dict[str, str] | None = None,
) -> dict[str, Any]:
    """校验一批导入行。

    soft=True：工程目录分析模式——按现场表头容错，不要求成品已入库；
    缺可执行工序/适用工序时给默认值并记 warning，便于动态识别关键字段。
    """
    ok: list[dict[str, Any]] = []
    errors: list[str] = []
    warnings: list[str] = []
    diagnostics: list[dict[str, Any]] = []
    sheet_code = _sheet_as_code(sheet)
    pmap = product_by_order or {}

    def _source_for(raw: dict[str, Any], parser_index: int) -> dict[str, Any]:
        source = raw.get(_XLSX_SOURCE_KEY)
        if isinstance(source, dict):
            return source
        summary: dict[str, Any] = {}
        for key, value in raw.items():
            if str(key).startswith("__") or _is_blank(value):
                continue
            summary[str(key)] = _jsonish_value(value)
            if len(summary) >= 8:
                break
        return {"sheet": sheet, "excelRow": None, "rawSummary": summary, "section": ""}

    def _make_diagnostic(
        raw: dict[str, Any], parser_index: int, *, code: str,
        severity: str, message: str, diagnostic_kind: str | None = None,
        classification: str | None = None,
    ) -> dict[str, Any]:
        source = _source_for(raw, parser_index)
        excel_row = source.get("excelRow")
        inferred_classification = classification
        if not inferred_classification:
            if code == "DUPLICATE_ALIAS_EMPTY_IGNORED":
                inferred_classification = "parser_noise"
            elif code in {"SECONDARY_SEQUENCE_HEADER", "EMBEDDED_SECTION_ROW"}:
                inferred_classification = "ignorable_section"
            else:
                inferred_classification = severity
        return {
            "sheet": source.get("sheet") or sheet or "",
            "excelRow": int(excel_row) if isinstance(excel_row, (int, float)) else None,
            "parserIndex": parser_index,
            "kind": diagnostic_kind or kind,
            "code": code,
            "severity": severity,
            "classification": inferred_classification,
            "message": message,
            "rawSummary": source.get("rawSummary") or {},
        }

    def _add_warning(raw: dict[str, Any], parser_index: int, code: str, message: str) -> None:
        warnings.append(f"第{parser_index}行：{message}")
        diagnostics.append(_make_diagnostic(
            raw, parser_index, code=code, severity="warning", message=message,
        ))

    def _blocking_code(raw: dict[str, Any], message: str) -> tuple[str, str]:
        if kind == "equipment":
            if _is_blank(raw.get("code")):
                return (
                    "EQUIPMENT_CODE_MISSING",
                    "设备编码缺失：财务编号、固定资产编号和厂内编号均为空",
                )
            return "EQUIPMENT_CAPABILITIES_MISSING", "设备缺少可执行工序"
        codes = {
            "orders": "ORDER_REQUIRED_FIELDS_MISSING",
            "materials": "MATERIAL_REQUIRED_FIELDS_MISSING",
            "calendar": "CALENDAR_REQUIRED_FIELDS_MISSING",
            "molds": "MOLD_REQUIRED_FIELDS_MISSING",
            "operations": "OPERATION_CODE_MISSING",
            "zones": "ZONE_CODE_MISSING",
            "routing": "ROUTING_REQUIRED_FIELDS_MISSING",
            "bom": "BOM_REQUIRED_FIELDS_MISSING",
        }
        return codes.get(kind, "ROW_VALIDATION_ERROR"), message

    def _seq_val(v: Any) -> int:
        if v is None or v == "":
            return 0
        try:
            return int(float(v))
        except (TypeError, ValueError):
            s = str(v).strip()
            digits = re.sub(r"\D", "", s)
            return int(digits) if digits else 0

    for i, raw in enumerate(rows, 1):
        source = _source_for(raw, i)
        parser_diagnostic = raw.get(_XLSX_DIAGNOSTIC_KEY)
        if isinstance(parser_diagnostic, dict):
            diagnostics.append(_make_diagnostic(
                raw,
                i,
                code=str(parser_diagnostic.get("code") or "PARSER_IGNORED_ROW"),
                severity=str(parser_diagnostic.get("severity") or "ignored"),
                message=str(parser_diagnostic.get("message") or "结构性行已忽略"),
                diagnostic_kind=str(parser_diagnostic.get("kind") or kind),
                classification=str(parser_diagnostic.get("classification") or "") or None,
            ))
            continue
        for normalization in source.get("normalizations") or []:
            if not isinstance(normalization, dict):
                continue
            diagnostics.append(_make_diagnostic(
                raw, i,
                code=str(normalization.get("code") or "DUPLICATE_ALIAS_NORMALIZED"),
                severity="warning",
                message=str(normalization.get("message") or "重复别名已规范化"),
            ))
        try:
            if kind == "orders":
                from server.aps_domain.orders import find_product_by_hint
                cust = str(raw.get("customerName") or "").strip()
                hint = str(raw.get("productCode") or raw.get("productName") or "").strip()
                # 料号形如 200042588[282000009 乘致电机] → 取括号内编码优先，否则整段
                m_bracket = re.search(r"\[([^\]]+)\]", hint)
                if m_bracket:
                    inner = m_bracket.group(1).strip()
                    code_part = re.split(r"\s+", inner, maxsplit=1)[0]
                    if code_part:
                        hint = code_part
                    if not str(raw.get("productName") or "").strip() and " " in inner:
                        raw = {**raw, "productName": inner.split(" ", 1)[1]}
                try:
                    qty = int(float(raw.get("quantity") or 0))
                except (TypeError, ValueError):
                    qty = 0
                due = _due_str(raw.get("deliveryDate"))
                if soft:
                    if not hint or qty <= 0 or not due:
                        raise ValueError("需产品(料号)/数量/交期(计划结束)")
                    if not cust:
                        cust = "现场项目"
                        _add_warning(
                        raw, i, "ORDER_CUSTOMER_DEFAULTED",
                        "无客户列，暂用「现场项目」（可用项目名称）",
                    )
                elif not cust or not hint or qty <= 0 or not due:
                    raise ValueError("需客户/产品/数量/交期")
                prod = find_product_by_hint(world, hint)
                if prod is None:
                    prod = next((m for m in world.get("flexMaterials", [])
                                 if m.get("type") == "FINISHED_PRODUCT"
                                 and (m["code"].lower() == hint.lower() or hint in m.get("name", ""))), None)
                if prod is None and not soft:
                    raise ValueError(f"找不到成品「{hint}」")
                row_out = {
                    "customerName": cust,
                    "customerLevel": str(raw.get("customerLevel") or "A").upper(),
                    "productId": (prod or {}).get("id"),
                    "productCode": (prod or {}).get("code") or hint,
                    "productName": (prod or {}).get("name") or str(raw.get("productName") or hint),
                    "quantity": qty, "deliveryDate": due,
                    "priority": int(float(raw.get("priority") or 5)),
                    "isRush": _truthy(raw.get("isRush")), "status": "CONFIRMED",
                }
                ono = str(raw.get("orderNo") or "").strip()
                if ono:
                    row_out["orderNo"] = ono
                ok.append(row_out)
            elif kind == "materials":
                code = str(raw.get("code") or raw.get("materialCode") or raw.get("productCode") or "").strip()
                name = str(raw.get("name") or raw.get("productName") or "").strip()
                if not code or not name:
                    raise ValueError("需编码与名称")
                type_map = {"成品": "FINISHED_PRODUCT", "半成品": "SEMI_FINISHED",
                            "原料": "RAW_MATERIAL", "原材料": "RAW_MATERIAL"}
                raw_type = str(raw.get("type") or "").strip()
                mtype = type_map.get(raw_type, raw_type.upper() if raw_type else "RAW_MATERIAL")
                if soft and mtype not in ("FINISHED_PRODUCT", "SEMI_FINISHED", "RAW_MATERIAL"):
                    mtype = "RAW_MATERIAL"
                ok.append({
                    "code": code, "name": name, "type": mtype,
                    "unit": str(raw.get("unit") or "件").strip('"') or "件",
                    "stock": float(raw.get("stock") or 0),
                    "inTransit": float(raw.get("inTransit") or 0),
                    "safetyStock": float(raw.get("safetyStock") or 0),
                    "procurementLeadTime": int(float(raw.get("procurementLeadTime") or 0)),
                    "spec": str(raw.get("spec") or ""),
                    "expectedArrivalDate": str(raw.get("expectedArrivalDate") or ""),
                    "sourcingType": str(raw.get("sourcingType") or ""),
                })
            elif kind == "calendar":
                shift_code = str(raw.get("shiftCode") or raw.get("code") or "").strip()
                start_time = str(raw.get("startTime") or "").strip()
                end_time = str(raw.get("endTime") or "").strip()
                if not shift_code or not re.fullmatch(r"\d{2}:\d{2}", start_time) or not re.fullmatch(r"\d{2}:\d{2}", end_time):
                    raise ValueError("需班次编码、开始时间与结束时间(HH:MM)")
                workdays: list[int] = []
                for value in _split_multi(raw.get("workdays")) or ["1", "2", "3", "4", "5", "6", "7"]:
                    try:
                        day = int(float(value))
                    except (TypeError, ValueError):
                        continue
                    if 1 <= day <= 7 and day not in workdays:
                        workdays.append(day)
                breaks: list[dict[str, str]] = []
                for value in _split_multi(raw.get("breaks")):
                    pair = re.split(r"\s*[-~～至]\s*", value, maxsplit=1)
                    if len(pair) == 2 and all(re.fullmatch(r"\d{2}:\d{2}", part) for part in pair):
                        breaks.append({"start": pair[0], "end": pair[1]})
                enabled_raw = raw.get("enabled")
                ok.append({
                    "shiftCode": shift_code,
                    "name": str(raw.get("name") or shift_code),
                    "startTime": start_time,
                    "endTime": end_time,
                    "breaks": breaks,
                    "workdays": workdays or [1, 2, 3, 4, 5, 6, 7],
                    "enabled": True if enabled_raw in (None, "") else _truthy(enabled_raw),
                    "demoOnly": _truthy(raw.get("demoOnly")),
                })
            elif kind == "equipment":
                code = str(raw.get("code") or "").strip()
                name = str(raw.get("name") or code).strip()
                caps = _split_multi(raw.get("capabilities"))
                op_std = _parse_op_std(raw.get("opStdTime"))
                if not op_std and caps:
                    op_std = {c: 1.0 for c in caps}
                if not caps and op_std:
                    caps = list(op_std.keys())
                if soft and code and not caps:
                    caps = ["GENERAL"]
                    op_std = {"GENERAL": 1.0}
                    _add_warning(
                        raw, i, "EQUIPMENT_CAPABILITIES_DEFAULTED",
                        "设备无「可执行工序」，暂记 GENERAL（可按工艺推断）",
                    )
                if not code or not caps:
                    raise ValueError("需设备编码与可执行工序")
                ok.append({
                    "code": code, "name": name, "capabilities": caps, "opStdTime": op_std,
                    "movable": _truthy(raw.get("movable")),
                    "moveTimeMin": float(raw.get("moveTimeMin") or 0),
                    "zone": str(raw.get("zone") or "ZONE-A"),
                    "adaptableMolds": _split_multi(raw.get("adaptableMolds")),
                    "availabilityRate": float(raw.get("availabilityRate") or 0.95),
                    "status": str(raw.get("status") or "RUNNING").upper(),
                })
            elif kind == "molds":
                code = str(raw.get("code") or "").strip()
                op = str(raw.get("operationCode") or raw.get("type") or "").strip()
                if soft and code and not op:
                    op = "GENERAL"
                    _add_warning(
                        raw, i, "MOLD_OPERATION_DEFAULTED",
                        "模具无适用工序，暂记 GENERAL",
                    )
                if not code or not op:
                    raise ValueError("需模具编码与适用工序")
                ok.append({
                    "code": code, "name": str(raw.get("name") or code),
                    "operationCode": op,
                    "adaptableEquipment": _split_multi(raw.get("adaptableEquipment")),
                    "lifeTotal": int(float(raw.get("lifeTotal") or 50000)),
                    "lifeUsed": int(float(raw.get("lifeUsed") or 0)),
                    "zone": str(raw.get("zone") or "ZONE-A"),
                    "changeoverMin": float(raw.get("changeoverMin") or 20),
                    "status": str(raw.get("status") or "AVAILABLE").upper(),
                })
            elif kind == "operations":
                code = str(raw.get("code") or raw.get("operationCode") or "").strip()
                if not code:
                    raise ValueError("需工序编码")
                ok.append({
                    "code": code, "name": str(raw.get("name") or code),
                    "isBottleneck": _truthy(raw.get("isBottleneck")),
                    "changeoverMin": float(raw.get("changeoverMin") or 0),
                })
            elif kind == "zones":
                code = str(raw.get("code") or "").strip()
                if not code:
                    raise ValueError("需区域编码")
                ok.append({"code": code, "name": str(raw.get("name") or code)})
            elif kind == "routing":
                pc = str(raw.get("productCode") or "").strip()
                if not pc and soft:
                    # sheet 名常为订单号 → 映射到料号；否则直接用 sheet
                    pc = pmap.get(sheet_code) or sheet_code
                seq = _seq_val(raw.get("seq"))
                op = str(raw.get("operationCode") or "").strip()
                if not op or seq <= 0:
                    raise ValueError("需产品编码/序号/工序")
                if not pc:
                    raise ValueError("需产品编码/序号/工序")
                std = raw.get("stdTimePerUnit")
                try:
                    std_f = float(std) if std not in (None, "") else None
                except (TypeError, ValueError):
                    std_f = None
                ok.append({
                    "productCode": pc,
                    "productName": str(raw.get("productName") or raw.get("name") or pc),
                    "seq": seq, "operationCode": op,
                    "operationName": str(raw.get("name") or raw.get("operationName") or op),
                    "requireMold": _truthy(raw.get("requireMold")),
                    "stdTimePerUnit": std_f,
                })
            elif kind == "bom":
                pc = str(raw.get("productCode") or "").strip()
                if not pc and soft:
                    ono = str(raw.get("orderNo") or sheet_code).strip()
                    pc = pmap.get(ono) or ono
                mc = str(raw.get("materialCode") or raw.get("code") or "").strip()
                qty = float(raw.get("quantity") or 0)
                if not pc or not mc or qty <= 0:
                    raise ValueError("需成品/子件/用量")
                ok.append({
                    "productCode": pc, "materialCode": mc, "quantity": qty,
                    "consumeOp": str(raw.get("consumeOp") or ""),
                    "isKey": _truthy(raw.get("isKey")),
                })
            else:
                raise ValueError(f"不支持的导入类型：{kind}")
        except Exception as exc:  # noqa: BLE001
            message = str(exc)
            errors.append(f"第{i}行：{message}")
            code, diagnostic_message = _blocking_code(raw, message)
            diagnostics.append(_make_diagnostic(
                raw, i, code=code, severity="blocking",
                message=diagnostic_message,
            ))
    # soft 模式下同类 warning 去重截断
    if warnings:
        uniq: list[str] = []
        seen_w: set[str] = set()
        for w in warnings:
            key = re.sub(r"第\d+行：", "第N行：", w)
            if key in seen_w:
                continue
            seen_w.add(key)
            uniq.append(w)
            if len(uniq) >= 5:
                break
        warnings = uniq
    counts = _diagnostic_counts(diagnostics)
    return {
        "kind": kind, "okRows": ok, "errors": errors, "warnings": warnings,
        "okCount": len(ok), "errorCount": counts["blocking"],
        "diagnostics": diagnostics, "diagnosticCounts": counts,
    }


def preview_file(
    filename: str,
    raw: bytes,
    world: World,
    *,
    soft: bool = False,
    product_by_order: dict[str, str] | None = None,
    kind_hint: str | None = None,
) -> dict[str, Any]:
    name = filename or "upload"
    lower = name.lower()
    batches = []
    if lower.endswith((".csv", ".txt")):
        headers, rows = _rows_from_csv(raw)
        kind = kind_hint or detect_kind(name, headers) or "materials"
        fmap = detect_field_map(kind, headers)
        batch = validate_batch(
            kind, rows, world, soft=soft, sheet=name, product_by_order=product_by_order)
        batches.append({
            "sheet": name, "physicalSheet": name,
            "kind": kind, "fieldMap": fmap, "headersRaw": headers, **batch,
        })
    elif lower.endswith((".xlsx", ".xlsm")):
        from server.importers.planning_workbook import preview_planning_workbook

        structured = preview_planning_workbook(name, raw)
        if structured is not None:
            return structured
        for sheet_data in _rows_from_xlsx(raw):
            kind = (
                kind_hint
                or sheet_data.get("kind")
                or detect_kind(f"{name}:{sheet_data['sheet']}", sheet_data["headersRaw"])
                or detect_kind(name, sheet_data["headers"])
                or detect_kind(name, sheet_data["headersRaw"])
                or "materials"
            )
            field_map = detect_field_map(kind, sheet_data["headersRaw"])
            physical_sheet = sheet_data.get("physicalSheet") or sheet_data["sheet"]
            batch = validate_batch(
                kind, sheet_data["rows"], world, soft=soft,
                sheet=physical_sheet, product_by_order=product_by_order,
            )
            batches.append({
                "sheet": sheet_data["sheet"],
                "physicalSheet": physical_sheet,
                "kind": kind,
                "fieldMap": field_map,
                "headersRaw": sheet_data["headersRaw"],
                **batch,
            })
    else:
        raise ValueError("仅支持 .xlsx / .csv")

    diagnostics = [
        diagnostic
        for batch in batches
        for diagnostic in (batch.get("diagnostics") or [])
    ]
    diagnostic_counts = _diagnostic_counts(diagnostics)
    total_ok = sum(batch["okCount"] for batch in batches)
    total_err = sum(batch["errorCount"] for batch in batches)
    return {
        "filename": name,
        "batches": batches,
        "totalOk": total_ok,
        "totalErrors": total_err,
        "canCommit": total_ok > 0,
        "diagnostics": diagnostics,
        "diagnosticCounts": diagnostic_counts,
        "diagnosticsVersion": 1,
    }


def confirmation_for_import_commit(preview: dict[str, Any]) -> tuple[str, list[str]]:
    title = f"导入入库 · {preview.get('filename', '')}"
    lines = [f"有效 {preview.get('totalOk', 0)} 行 · 错误 {preview.get('totalErrors', 0)} 行（错误行不入库）"]
    for b in preview.get("batches", []):
        lines.append(f"· [{b.get('sheet')}] {b['kind']} 通过 {b['okCount']} / 失败 {b['errorCount']}")
    lines.append("批准后写入主干（P2）；执行前自动建档可回滚")
    return title, lines


def apply_import_commit(world: World, next_id, batches: list[dict[str, Any]]) -> dict[str, Any]:
    from server.importers.planning_workbook import apply_planning_batches

    profile_batches = [b for b in batches if b.get("sourceProfile")]
    if profile_batches:
        if len(profile_batches) != len(batches):
            raise ValueError("完整工作簿与其他导入批次不能混合确认，请分别导入")
        return apply_planning_batches(world, profile_batches)
    from server.aps_domain.masterdata import apply_master_action
    from server.aps_domain.orders import apply_order_action
    from server.state.seed import ensure_flex_seed

    ensure_flex_seed(world)
    summary: dict[str, int] = {}

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

    for batch in batches:
        kind = batch["kind"]
        for row in batch.get("okRows") or []:
            if kind == "orders":
                items = world.setdefault("flexOrders", [])
                order_no = row.get("orderNo")
                existing = next((item for item in items if order_no and item.get("orderNo") == order_no), None)
                if row.get("productId") and existing is None and not order_no:
                    apply_order_action(world, next_id, "order.upsert", row)
                    summary["salesOrders"] = summary.get("salesOrders", 0) + 1
                else:
                    values = {
                        "orderNo": order_no,
                        "productCode": row.get("productCode"), "quantity": row["quantity"],
                        "dueDate": row["deliveryDate"], "priority": row.get("priority", 5),
                        "customerName": row.get("customerName") or "现场客户",
                        "customerLevel": row.get("customerLevel") or "B",
                        "isRush": bool(row.get("isRush")),
                        "status": "RELEASED",
                    }
                    if existing:
                        existing.update(values)
                    else:
                        mid = _next_table_id("flexOrders")
                        items.append({"id": mid, **values, "orderNo": order_no or f"FO-{mid:04d}"})
                    summary["flexOrders"] = summary.get("flexOrders", 0) + 1
            elif kind == "materials":
                existing_material = next(
                    (item for item in world.setdefault("materials", []) if item.get("code") == row.get("code")),
                    None,
                )
                if existing_material:
                    master_row = {**row, "id": existing_material["id"]}
                else:
                    # 订单先入库时 sync_flex_orders_to_sales 可能已直接建成品，
                    # 外部发号器仍停在 0，直接 create 会让原材料与成品 id 冲突。
                    master_row = {**row, "newId": _next_table_id("materials")}
                apply_master_action(world, next_id, "master.material.upsert", master_row)
                fm = world.setdefault("flexMaterials", [])
                ex = next((m for m in fm if m["code"] == row["code"]), None)
                if ex:
                    ex.update({k: row[k] for k in (
                        "name", "type", "unit", "stock", "inTransit", "safetyStock",
                        "procurementLeadTime", "expectedArrivalDate", "sourcingType",
                    ) if k in row})
                else:
                    fm.append({"id": _next_table_id("flexMaterials"),
                               **{k: row.get(k) for k in
                                  ("code", "name", "type", "unit", "stock", "inTransit",
                                   "safetyStock", "procurementLeadTime", "expectedArrivalDate", "sourcingType")}})
                if existing_material is None:
                    summary["materials"] = summary.get("materials", 0) + 1
            elif kind == "calendar":
                items = world.setdefault("flexCalendar", [])
                ex = next((item for item in items if item.get("shiftCode") == row.get("shiftCode")), None)
                if ex:
                    ex.update(row)
                else:
                    items.append(dict(row))
                summary["flexCalendar"] = summary.get("flexCalendar", 0) + 1
            elif kind == "equipment":
                items = world.setdefault("flexEquipment", [])
                ex = next((e for e in items if e["code"] == row["code"]), None)
                if ex:
                    ex.update(row)
                else:
                    items.append({"id": _next_table_id("flexEquipment"), **row})
                summary["flexEquipment"] = summary.get("flexEquipment", 0) + 1
            elif kind == "molds":
                items = world.setdefault("flexMolds", [])
                ex = next((e for e in items if e["code"] == row["code"]), None)
                if ex:
                    ex.update(row)
                else:
                    items.append({"id": _next_table_id("flexMolds"), **row})
                summary["flexMolds"] = summary.get("flexMolds", 0) + 1
            elif kind == "operations":
                items = world.setdefault("flexOperations", [])
                ex = next((e for e in items if e["code"] == row["code"]), None)
                if ex:
                    ex.update(row)
                else:
                    items.append(row)
                summary["flexOperations"] = summary.get("flexOperations", 0) + 1
            elif kind == "zones":
                items = world.setdefault("flexZones", [])
                ex = next((e for e in items if e["code"] == row["code"]), None)
                if ex:
                    ex.update(row)
                else:
                    items.append(row)
                summary["flexZones"] = summary.get("flexZones", 0) + 1
            elif kind == "routing":
                items = world.setdefault("flexRoutings", [])
                items[:] = [s for s in items
                            if not (s.get("productCode") == row["productCode"] and s.get("seq") == row["seq"])]
                items.append(dict(row))
                summary["flexRoutings"] = summary.get("flexRoutings", 0) + 1
            elif kind == "bom":
                items = world.setdefault("flexBom", [])
                items[:] = [s for s in items
                            if not (s.get("productCode") == row["productCode"]
                                    and s.get("materialCode") == row["materialCode"])]
                items.append(dict(row))
                summary["flexBom"] = summary.get("flexBom", 0) + 1
    # 通用表格导入首先写入 flex* 结构；MRP、订单详情和固定轨排产读取经典主数据。
    # 批次全部落地后统一做幂等投影，避免出现“分析卡显示 BOM/工艺，但 MRP 只能看到根成品”的双轨漂移。
    if any(summary.get(key, 0) for key in (
        "materials", "flexEquipment", "flexOperations", "flexRoutings", "flexBom",
    )):
        from server.importers.sql_pack import sync_flex_to_classic_master

        sync_flex_to_classic_master(world)
    if summary.get("flexOrders", 0):
        from server.aps_domain.orders import sync_flex_orders_to_sales

        sync_flex_orders_to_sales(world)
    return {"summary": summary, "total": sum(summary.values())}