fix: 统一mock流水可识别身份证来源

lsfx-mock-server/services/statement_service.py

@@ -1,9 +1,13 @@
-from utils.response_builder import ResponseBuilder
-from typing import Dict, Union, List
-import random
-from datetime import datetime, timedelta
-import uuid
+from typing import Dict, List, Union
 import logging
+import random
+import uuid
+from datetime import datetime, timedelta
+
+from services.statement_rule_samples import (
+    IDENTITY_CARD_POOL,
+    build_large_transaction_seed_statements,
+)
 
 # 配置日志
 logging.basicConfig(level=logging.INFO)
@@ -17,102 +21,75 @@ class StatementService:
         # 缓存：logId -> (statements_list, total_count)
         self._cache: Dict[int, tuple] = {}
         self.file_service = file_service
-        # 配置日志级别为 INFO
-        logger.info(f"StatementService initialized with empty cache")
+        logger.info("StatementService initialized with empty cache")
 
     def _resolve_primary_binding(self, log_id: int) -> tuple:
-        """优先从 FileService 读取真实主绑定，不存在时再走 fallback。"""
+        """优先从 FileService 读取真实主绑定，不存在时再走 deterministic fallback。"""
         if self.file_service is not None:
             record = self.file_service.get_file_record(log_id)
             if record is not None:
                 return record.primary_enterprise_name, record.primary_account_no
 
-        return "张传伟", f"{random.randint(100000000000000, 999999999999999)}"
+        rng = random.Random(f"binding:{log_id}")
+        return "张传伟", f"{rng.randint(100000000000000, 999999999999999)}"
 
     def _generate_random_statement(
         self,
-        index: int,
         group_id: int,
         log_id: int,
         primary_enterprise_name: str,
         primary_account_no: str,
+        rng: random.Random,
     ) -> Dict:
-        """生成单条随机流水记录
+        """生成单条随机噪声流水记录。"""
+        reference_now = datetime(2026, 3, 18, 9, 0, 0)
+        days_ago = rng.randint(0, 365)
+        trx_datetime = reference_now - timedelta(days=days_ago, minutes=rng.randint(0, 1439))
+        trans_amount = round(rng.uniform(10, 10000), 2)
 
-        Args:
-            index: 流水序号
-            group_id: 项目ID
-            log_id: 文件ID
-            primary_enterprise_name: 本方主体名称
-            primary_account_no: 本方账号
-
-        Returns:
-            单条流水记录字典
-        """
-        # 随机生成交易日期（最近1年内）
-        days_ago = random.randint(0, 365)
-        trx_datetime = datetime.now() - timedelta(days=days_ago)
-        trx_date = trx_datetime.strftime("%Y-%m-%d %H:%M:%S")
-        accounting_date = trx_datetime.strftime("%Y-%m-%d")
-        accounting_date_id = int(trx_datetime.strftime("%Y%m%d"))
-
-        # 生成创建日期（格式：YYYY-MM-DD HH:MM:SS）
-        create_date = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-
-        # 随机生成交易金额
-        trans_amount = round(random.uniform(10, 10000), 2)
-
-        # 随机决定是收入还是支出
-        if random.random() > 0.5:
-            # 支出
+        if rng.random() > 0.5:
             dr_amount = trans_amount
-            cr_amount = 0
+            cr_amount = 0.0
             trans_flag = "P"
         else:
-            # 收入
             cr_amount = trans_amount
-            dr_amount = 0
+            dr_amount = 0.0
             trans_flag = "R"
 
-        # 随机余额
-        balance_amount = round(random.uniform(1000, 50000), 2)
-
-        # 随机客户信息
-        customers = ["小店", "支付宝", "微信支付", "财付通", "美团", "京东", "淘宝", "银行转账"]
-        customer_name = random.choice(customers)
-        customer_account = str(random.randint(100000000, 999999999))
-
-        # 随机交易描述
-        memos = [
-            f"消费_{customer_name}",
-            f"转账_{customer_name}",
-            f"收款_{customer_name}",
-            f"支付_{customer_name}",
-            f"退款_{customer_name}",
-        ]
-        user_memo = random.choice(memos)
+        customer_name = rng.choice(
+            ["小店", "支付宝", "微信支付", "财付通", "美团", "京东", "淘宝", "银行转账"]
+        )
+        user_memo = rng.choice(
+            [
+                f"消费_{customer_name}",
+                f"转账_{customer_name}",
+                f"收款_{customer_name}",
+                f"支付_{customer_name}",
+                f"退款_{customer_name}",
+            ]
+        )
 
         return {
             "accountId": 0,
             "accountMaskNo": primary_account_no,
-            "accountingDate": accounting_date,
-            "accountingDateId": accounting_date_id,
+            "accountingDate": trx_datetime.strftime("%Y-%m-%d"),
+            "accountingDateId": int(trx_datetime.strftime("%Y%m%d")),
             "archivingFlag": 0,
             "attachments": 0,
-            "balanceAmount": balance_amount,
+            "balanceAmount": round(rng.uniform(1000, 50000), 2),
             "bank": "ZJRCU",
             "bankComments": "",
-            "bankStatementId": 12847662 + index,
-            "bankTrxNumber": uuid.uuid4().hex,
+            "bankStatementId": 0,
+            "bankTrxNumber": "",
             "batchId": log_id,
             "cashType": "1",
             "commentsNum": 0,
             "crAmount": cr_amount,
-            "createDate": create_date,
+            "createDate": reference_now.strftime("%Y-%m-%d %H:%M:%S"),
             "createdBy": "902001",
-            "cretNo": "230902199012261247",
+            "cretNo": rng.choice(IDENTITY_CARD_POOL),
             "currency": "CNY",
-            "customerAccountMaskNo": customer_account,
+            "customerAccountMaskNo": str(rng.randint(100000000, 999999999)),
             "customerBank": "",
             "customerId": -1,
             "customerName": customer_name,
@@ -138,36 +115,54 @@ class StatementService:
             "transformDrAmount": 0,
             "transfromBalanceAmount": 0,
             "trxBalance": 0,
-            "trxDate": trx_date,
-            "uploadSequnceNumber": index + 1,
-            "userMemo": user_memo
+            "trxDate": trx_datetime.strftime("%Y-%m-%d %H:%M:%S"),
+            "uploadSequnceNumber": 0,
+            "userMemo": user_memo,
         }
 
-
+    def _assign_statement_ids(self, statements: List[Dict], group_id: int, log_id: int) -> List[Dict]:
+        """为样本与噪声流水统一补齐稳定的流水标识。"""
+        assigned: List[Dict] = []
+        base_id = log_id * 100000
+        for index, statement in enumerate(statements, start=1):
+            item = dict(statement)
+            item["groupId"] = group_id
+            item["batchId"] = log_id
+            item["bankStatementId"] = base_id + index
+            item["bankTrxNumber"] = uuid.uuid5(
+                uuid.NAMESPACE_DNS, f"lsfx-mock-{log_id}-{index}"
+            ).hex
+            item["uploadSequnceNumber"] = index
+            item["transAmount"] = round(item.get("drAmount", 0) + item.get("crAmount", 0), 2)
+            assigned.append(item)
+        return assigned
 
     def _generate_statements(self, group_id: int, log_id: int, count: int) -> List[Dict]:
-        """生成指定数量的流水记录
-
-        Args:
-            group_id: 项目ID
-            log_id: 文件ID
-            count: 生成数量
-
-        Returns:
-            流水记录列表
-        """
+        """生成指定数量的流水记录。"""
         primary_enterprise_name, primary_account_no = self._resolve_primary_binding(log_id)
-        statements = []
-        for i in range(count):
+        rng = random.Random(f"statement:{log_id}")
+        seeded_statements = build_large_transaction_seed_statements(
+            group_id=group_id,
+            log_id=log_id,
+            primary_enterprise_name=primary_enterprise_name,
+            primary_account_no=primary_account_no,
+        )
+
+        total_count = max(count, len(seeded_statements))
+        statements = list(seeded_statements)
+        for _ in range(total_count - len(seeded_statements)):
             statements.append(
                 self._generate_random_statement(
-                    i,
                     group_id,
                     log_id,
                     primary_enterprise_name,
                     primary_account_no,
+                    rng,
                 )
             )
+
+        statements = self._assign_statement_ids(statements, group_id, log_id)
+        rng.shuffle(statements)
         return statements
 
     def _apply_primary_binding(
@@ -182,15 +177,7 @@ class StatementService:
             statement["accountMaskNo"] = primary_account_no
 
     def get_bank_statement(self, request: Union[Dict, object]) -> Dict:
-        """获取银行流水列表
-
-        Args:
-            request: 获取银行流水请求（可以是字典或对象）
-
-        Returns:
-            银行流水响应字典
-        """
-        # 支持 dict 或对象
+        """获取银行流水列表。"""
         if isinstance(request, dict):
             group_id = request.get("groupId", 1000)
             log_id = request.get("logId", 10000)
@@ -202,25 +189,16 @@ class StatementService:
             page_now = request.pageNow
             page_size = request.pageSize
 
-        # 检查缓存中是否已有该logId的数据
         if log_id not in self._cache:
-            # 随机生成总条数（1200-1500之间）
-            total_count = random.randint(1200, 1500)
-            # 生成所有流水记录
+            total_rng = random.Random(f"total:{log_id}")
+            total_count = total_rng.randint(1200, 1500)
             all_statements = self._generate_statements(group_id, log_id, total_count)
-            # 存入缓存
             self._cache[log_id] = (all_statements, total_count)
 
-        # 从缓存获取数据
         all_statements, total_count = self._cache[log_id]
         primary_enterprise_name, primary_account_no = self._resolve_primary_binding(log_id)
-        self._apply_primary_binding(
-            all_statements,
-            primary_enterprise_name,
-            primary_account_no,
-        )
+        self._apply_primary_binding(all_statements, primary_enterprise_name, primary_account_no)
 
-        # 模拟分页
         start = (page_now - 1) * page_size
         end = start + page_size
         page_data = all_statements[start:end]