张胜宇
5d0becb67d
一、客服 Agent 智能增强（正面回应"不智能、动不动就转人工"）
- 决策链由 2 个出口扩到 5 个：E1 澄清 / E2 计算型 / E3 知识直返 / E4 证据约束生成 / E5 分级回退
- 转人工从"默认动作"降为最后一档 E5c，只保留 4 类白名单：
P0 反诈 / P1 账户与个人数据 / P2 写操作与争议 / 用户明确要求人工
- 46 条金标实测（修复前 → 修复后）：
转人工率 43.5% → 10.9%；出口准确率 45.7% → 100%；事实正确率 69.6% → 100%
禁忌违反 1 → 0；档位越权 / 无出处数字 / 误拒 四项零容忍全 0
- 安全不变量 INV-1~INV-5；零容忍规则未删，改的是挂载点
（输出侧字面黑名单 → 检索层档位隔离 + 判定层合规词表 + 输出守护）
二、知识库：档位单点化与物理隔离
- 新增 app/core/knowledge_tier.py 作为档位规则唯一落点（G-03），
knowledge_contracts.py 原定义块改为显式再导出（X as X，非副本）
- 档位过滤由 bool 默认值（fail-open）改为 tiers 必填集合（缺参即 TypeError）
- Milvus 侧四集合按 visibility 分区键物理隔离；双 schema 收敛为一套
- 新增 app/core/actor.py：访客三元组与匿名判定的唯一构造/判定点（G-01/G-01b）
- 新增 app/core/fund_fee_rules.py：费率计算纯函数
三、前端入参边界对齐（本轮 W11 新修，4 处"校验宽于存储"）
- message 加 max_length=8000（与浮窗 widget.js 的 maxlength 一致）
- session_id 加 1—64；idempotency_key 上限 128 → 64（对齐列宽 String(64)）
- feedback_type 加 max_length=32（对齐列宽 String(32)）
- 8 条路径参数补 min_length=1 + max_length=64 + 字符集正则
（{session_id} / {run_id} / {handover_id}）
- 改前超限值会落到 MySQL 才失败（500）；改后一律 422 AGENT_INPUT_INVALID + 字段级定位
- 新增 tests/unit/api/test_frontend_boundaries.py（33 例），含"端点表 ↔ OpenAPI 全量对照"
四、投顾模块整体清除（D4.4 / D4.5）
- 删除投顾相关 controller / schema / model / repository / service 及门户页面
- tools/portal_api_check.py 同步作废 AD003/AD005/AD011/A047 四条用例与 advisor_t 登录
（端点与账号均已不存在，此前稳定报 3 条假红）
五、验证（提交前实测）
- pytest -q：1856 passed / 2 skipped / 0 failed
- ruff check app tools tests：19（= 基线）；mypy app：2（= 基线）
- 前端接口契约体检 portal_api_check.py：38 项，通过 34，失败 0，跳过 4
- 全链路冒烟 e2e_smoke_test.py --read-only：31/31
- HTTP 全链路探针 http_probe.py：11/11 succeeded
- 跨文档一致性 _consistency.py：GATE PASS
- 真机边界复验 12 条：12/12 符合预期
六、纪律与文档
- 可改文件白名单 A-09（docs/46）与底座会签申请单 A-10（docs/47，组 1—组 4 全部受理）
- 零 DDL：未新增/修改任何表结构，89 张业务表与基线一致
- 证据留痕：docs/evidence/**（含 46 条金标 score、快照、清除与重建记录）
- 未提交（刻意排除，见提交说明）：仓库内 客服agent/ 与 开发文档/ 是 2026-09-16 前的
过期副本（Todolist 440 行 vs 权威 D2.1 1167 行），权威正本在仓库外；
_chunks_report.txt 是 tools/build_knowledge_chunks.py 生成的本地产物
145 lines
7.4 KiB
Python
145 lines
7.4 KiB
Python
from functools import lru_cache
|
||
|
||
from dotenv import load_dotenv
|
||
from pydantic import Field
|
||
from pydantic_settings import BaseSettings, SettingsConfigDict
|
||
|
||
# 把 `.env` 显式注入进程环境变量。pydantic-settings 只把 `.env` 读进 Settings 对象，
|
||
# **不会**写 `os.environ`；而模型密钥解析（`EnvironmentSecretResolver` 要求 secret_ref
|
||
# 形如 `env:VAR`）读的正是 `os.getenv`。两者不打通就会出现"按文档在 .env 里配好密钥、
|
||
# 运行时却读不到"，且症状伪装成"没有可用的模型端点"，排查方向被带偏。
|
||
# `override=False`：真实进程环境变量优先，`.env` 只作补充（生产用容器注入时不会被覆盖）。
|
||
load_dotenv(override=False)
|
||
|
||
|
||
class Settings(BaseSettings):
|
||
app_env: str = "local"
|
||
app_name: str = "jr-agent-platform"
|
||
log_level: str = "INFO"
|
||
timezone: str = "Asia/Shanghai"
|
||
cors_allowed_origins: str = "http://127.0.0.1:5173,http://localhost:5173"
|
||
jwt_issuer: str
|
||
jwt_audience: str
|
||
jwt_algorithm: str = "RS256"
|
||
# 默认指向**开发专用**密钥（tools/generate_jwt_keys.py 生成，config/jwt/ 不进版本库）。
|
||
# 生产环境必须用环境变量覆盖为生产机上单独生成的那一套，不要复用开发密钥。
|
||
jwt_private_key_path: str = "config/jwt/dev/jwt-private.pem"
|
||
jwt_public_key_path: str = "config/jwt/dev/jwt-public.pem"
|
||
jwt_clock_skew_seconds: int = Field(default=30, ge=0)
|
||
visitor_token_ttl_seconds: int = Field(default=900, ge=60, le=3600)
|
||
mysql_dsn: str
|
||
mysql_pool_size: int = Field(default=5, ge=1)
|
||
mysql_max_overflow: int = Field(default=10, ge=0)
|
||
mysql_pool_pre_ping: bool = False
|
||
mysql_tx_isolation: str = "READ COMMITTED"
|
||
redis_url: str
|
||
redis_connect_timeout_seconds: float = Field(default=2, gt=0)
|
||
# 限流（文档 §3.5 的 429 / §3.6 的 RATE_LIMITED）。默认值刻意宽松：
|
||
# 每用户每接口每分钟 600 次（10 QPS）。取值依据是"本地验收与集成测试的量级"——
|
||
# 两个验收脚本与 tests/integration 里同一身份对同一接口最多几十次调用，
|
||
# 600 留了一个数量级余量，绝不可能误伤；同时 10 QPS 已足以拦住单个客户端
|
||
# 把底座打爆（正常业务交互远低于 1 QPS）。生产按真实容量收紧即可，无需改代码。
|
||
rate_limit_enabled: bool = True
|
||
rate_limit_window_seconds: int = Field(default=60, gt=0)
|
||
rate_limit_max_requests: int = Field(default=600, ge=1)
|
||
rate_limit_key_prefix: str = "jr:rate_limit"
|
||
message_broker_url: str = ""
|
||
message_broker_run_topic: str = "agent.run"
|
||
message_broker_outbox_topic: str = "agent.outbox"
|
||
message_broker_dlq_topic: str = "agent.dlq"
|
||
milvus_uri: str
|
||
milvus_local_uri: str = ""
|
||
milvus_token: str = ""
|
||
# 长期记忆向量集合名**刻意不做成配置项**：它是写（投影）、读（召回）、删（清理）
|
||
# 三侧共用的代码级契约，集合 schema 与向量维度（1024）也由代码定义
|
||
# （`app/infrastructure/milvus_profile_projection.py` 的 `PROFILE_COLLECTION`）。
|
||
# 这里原本有个 `milvus_collection`（`jr_memory`），写路径却从没读过它 ——
|
||
# 只有读/删两侧读，于是"写进 A、从 B 查、从 B 删"，语义召回恒空、清理恒报成功却不删。
|
||
# 一个只在契约一侧生效的配置项比没有配置项更危险，故直接删除。
|
||
neo4j_uri: str
|
||
neo4j_database: str = "neo4j"
|
||
neo4j_username: str = "neo4j"
|
||
neo4j_password: str = ""
|
||
model_router_config_ref: str = "local"
|
||
model_default_endpoint: str = ""
|
||
model_fallback_endpoint: str = ""
|
||
knowledge_embedding_endpoint_code: str = ""
|
||
knowledge_embedding_timeout_ms: int = Field(default=15000, gt=0)
|
||
sse_heartbeat_seconds: int = Field(default=15, gt=0)
|
||
sse_chunk_characters: int = Field(default=256, ge=1, le=4096)
|
||
sse_max_connection_seconds: int = Field(default=300, gt=0)
|
||
worker_lease_seconds: int = Field(default=60, gt=0)
|
||
worker_retry_limit: int = Field(default=3, ge=0)
|
||
worker_poll_seconds: float = Field(default=1, gt=0)
|
||
offsite_mailbox: str = "yuan80818843@163.com"
|
||
offsite_allowed_senders: tuple[str, ...] = ("15273589815@163.com",)
|
||
offsite_risk_receiver_id: str = ""
|
||
offsite_settlement_receiver_id: str = ""
|
||
offsite_mail_return_receiver: str = "yuan80818843@163.com"
|
||
offsite_max_retry_count: int = Field(default=3, ge=0)
|
||
offsite_imap_enabled: bool = False
|
||
offsite_imap_host: str = ""
|
||
offsite_imap_port: int = Field(default=993, gt=0)
|
||
offsite_imap_username: str = ""
|
||
offsite_imap_password: str = ""
|
||
offsite_imap_use_ssl: bool = True
|
||
offsite_imap_idle_enabled: bool = False
|
||
offsite_imap_idle_timeout_seconds: float = Field(default=120, gt=0)
|
||
offsite_mail_worker_enabled: bool = False
|
||
offsite_mail_worker_batch_size: int = Field(default=20, ge=1, le=100)
|
||
offsite_worker_user_id: str = ""
|
||
offsite_notification_sending_timeout_seconds: int = Field(default=900, gt=0)
|
||
offsite_mail_storage_dir: str = "data/offsite_mail"
|
||
offsite_ocr_enabled: bool = False
|
||
offsite_ocr_timeout_seconds: float = Field(default=30, gt=0)
|
||
offsite_aliyun_ocr_endpoint: str = "https://docmind-api.cn-hangzhou.aliyuncs.com"
|
||
offsite_aliyun_region_id: str = "cn-hangzhou"
|
||
offsite_aliyun_poll_interval_seconds: float = Field(default=1, gt=0)
|
||
offsite_aliyun_access_key_id: str = ""
|
||
offsite_aliyun_access_key_secret: str = ""
|
||
offsite_deepseek_enabled: bool = False
|
||
offsite_deepseek_base_url: str = "https://api.deepseek.com"
|
||
offsite_deepseek_api_key: str = ""
|
||
offsite_deepseek_model: str = "deepseek-v4-flash"
|
||
offsite_deepseek_timeout_seconds: float = Field(default=30, gt=0)
|
||
offsite_smtp_enabled: bool = False
|
||
offsite_smtp_dry_run: bool = True
|
||
offsite_smtp_host: str = ""
|
||
offsite_smtp_port: int = Field(default=465, gt=0)
|
||
offsite_smtp_username: str = ""
|
||
offsite_smtp_password: str = ""
|
||
offsite_smtp_sender: str = ""
|
||
offsite_smtp_use_ssl: bool = True
|
||
offsite_smtp_timeout_seconds: float = Field(default=30, gt=0)
|
||
promotion_material_storage_dir: str = "data/promotion_materials"
|
||
promotion_pdf_enabled: bool = False
|
||
promotion_pdf_converter_path: str = ""
|
||
promotion_max_photo_size_bytes: int = Field(default=10 * 1024 * 1024, gt=0)
|
||
promotion_max_performance_file_size_bytes: int = Field(default=20 * 1024 * 1024, gt=0)
|
||
promotion_image_enabled: bool = False
|
||
promotion_image_base_url: str = "https://dashscope.aliyuncs.com"
|
||
promotion_image_model: str = "wan2.2-t2i-flash"
|
||
promotion_image_timeout_seconds: float = Field(default=90, gt=0)
|
||
promotion_image_poll_interval_seconds: float = Field(default=2, gt=0)
|
||
|
||
risk_scan_schedule_enabled: bool = False
|
||
risk_scan_interval_minutes: int = Field(default=5, ge=1, le=1440)
|
||
risk_scan_run_immediately: bool = False
|
||
risk_scan_retry_limit: int = Field(default=2, ge=0, le=5)
|
||
risk_scan_poll_seconds: float = Field(default=30, gt=0)
|
||
risk_alert_mail_enabled: bool = False
|
||
risk_alert_mail_dry_run: bool = True
|
||
risk_alert_mail_recipients: str = ""
|
||
|
||
@property
|
||
def resolved_milvus_uri(self) -> str:
|
||
"""本地开发优先使用 Lite 文件；部署环境使用标准 Milvus URI。"""
|
||
return self.milvus_local_uri or self.milvus_uri
|
||
|
||
model_config = SettingsConfigDict(env_file=".env", env_file_encoding="utf-8", extra="ignore")
|
||
|
||
|
||
@lru_cache(maxsize=1)
|
||
def get_settings() -> Settings:
|
||
return Settings() # type: ignore[call-arg]
|