commit eaed17c36281cb31eb4f35c59d346ce6fd6430ed
Author: zuowei1216 <12206728+zuowei1216@user.noreply.gitee.com>
Date:   Mon May 25 19:35:22 2026 +0800

    feat: 开版订单 AI 解析服务
    
    - 豆包 AI 解析企微消息 → 开版订单字段
    - 字典枚举缓存（每日刷新，11 个分组 + 关联流程）
    - 布料精确匹配，找不到自动新增
    - 客户精确匹配，未找到返回建议
    - 企微消息代理（解决前端 CORS）
    - 支持 tags 写入（记录解析次数）
    - 单条消息图片获取接口

diff --git a/.env.example b/.env.example
new file mode 100644
index 0000000..0a2b4cc
--- /dev/null
+++ b/.env.example
@@ -0,0 +1,15 @@
+# 豆包 AI 配置
+AI_API_KEY=your-doubao-api-key
+AI_BASE_URL=https://ark.cn-beijing.volces.com/api/v3
+AI_MODEL=doubao-seed-2-0-mini-260428
+
+# ERP 后端配置（用于查询字典/客户/布料）
+ERP_BASE_URL=https://yuwenerp.yuwen.cloud
+ERP_USERNAME=your-username
+ERP_PASSWORD=your-password
+
+# 服务端口
+PORT=5050
+
+# 调试模式（生产环境改为 false）
+DEBUG=true
diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..b730881
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,22 @@
+# 环境变量（含密钥，不提交）
+.env
+.env.local
+.env.production
+.env.test
+
+# Python
+__pycache__/
+*.pyc
+*.pyo
+*.pyd
+.Python
+*.egg-info/
+dist/
+build/
+venv/
+.venv/
+
+# 测试/调试文件
+bench.py
+debug_api.py
+test_*.py
diff --git a/app.py b/app.py
new file mode 100644
index 0000000..3f4a45d
--- /dev/null
+++ b/app.py
@@ -0,0 +1,597 @@
+"""
+开版订单 AI 解析服务
+Flask + 豆包 API（/api/v3/responses 格式）
+
+功能：
+1. 代理企微消息接口（解决 CORS）
+2. AI 解析消息内容 → 开版订单字段
+3. 字典枚举值缓存（启动时拉取，每日刷新）
+4. 布料/客户智能匹配：搜索 → 不存在则自动新增
+
+启动：
+    python app.py
+"""
+
+import os
+import json
+import logging
+import threading
+import time
+import requests
+from datetime import datetime, timedelta
+from flask import Flask, request, jsonify
+from flask_cors import CORS
+from dotenv import load_dotenv
+
+load_dotenv()
+
+# ─── 日志 ─────────────────────────────────────────────────────────────────────
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s [%(levelname)s] %(message)s",
+)
+logger = logging.getLogger(__name__)
+
+# ─── Flask ────────────────────────────────────────────────────────────────────
+app = Flask(__name__)
+CORS(app, origins="*")
+
+# ─── 豆包 AI 配置 ─────────────────────────────────────────────────────────────
+AI_API_KEY  = os.getenv("AI_API_KEY", "")
+AI_BASE_URL = os.getenv("AI_BASE_URL", "https://ark.cn-beijing.volces.com/api/v3")
+AI_MODEL    = os.getenv("AI_MODEL", "doubao-seed-2-0-mini-260428")
+DOUBAO_URL  = f"{AI_BASE_URL}/responses"
+DOUBAO_HEADERS = {
+    "Authorization": f"Bearer {AI_API_KEY}",
+    "Content-Type": "application/json",
+}
+
+# ─── 企微消息服务配置 ─────────────────────────────────────────────────────────
+WECHAT_API_BASE = "https://test.ruicaiyinhua.online"
+WECHAT_AUTH     = "hophopkk"
+WORK_ORDER_ROOM = "wrW_mXZwAAaGwXBWvzbiAfEjy3ZM7ong"
+
+# ─── ERP 后端配置 ─────────────────────────────────────────────────────────────
+ERP_BASE_URL  = os.getenv("ERP_BASE_URL", "https://yuwenerp.yuwen.cloud")
+ERP_USERNAME  = os.getenv("ERP_USERNAME", "jimi")
+ERP_PASSWORD  = os.getenv("ERP_PASSWORD", "zuowei1216")
+
+# 需要全量缓存的小字典分组（用于 AI prompt 约束）
+SMALL_DICT_GROUPS = [
+    "起版情况", "紧急程度", "做货方式", "开版方式",
+    "地区", "画图评级", "调色评级", "套样评级",
+    "难度评级", "布源", "幅宽",
+]
+
+# ─── ERP Token 管理 ──────────────────────────────────────────────────────────
+class ErpAuth:
+    def __init__(self):
+        self.access_token = None
+        self.refresh_token = None
+        self.expires_at = 0
+        self._lock = threading.Lock()
+
+    def login(self):
+        with self._lock:
+            try:
+                r = requests.post(
+                    f"{ERP_BASE_URL}/api/auth/login/",
+                    json={"username": ERP_USERNAME, "password": ERP_PASSWORD},
+                    timeout=15,
+                )
+                r.raise_for_status()
+                data = r.json()
+                self.access_token  = data["access"]
+                self.refresh_token = data.get("refresh")
+                # JWT 默认 7 天，提前 1 小时刷新
+                self.expires_at = time.time() + 6 * 24 * 3600
+                logger.info("ERP 登录成功")
+                return self.access_token
+            except Exception as e:
+                logger.error(f"ERP 登录失败: {e}")
+                raise
+
+    def get_token(self):
+        if not self.access_token or time.time() >= self.expires_at:
+            return self.login()
+        return self.access_token
+
+    def headers(self):
+        return {"Authorization": f"Bearer {self.get_token()}"}
+
+erp_auth = ErpAuth()
+
+
+def erp_get(path: str, params: dict = None):
+    """GET ERP 接口，自动处理 401"""
+    url = f"{ERP_BASE_URL}{path}"
+    r = requests.get(url, headers=erp_auth.headers(), params=params or {}, timeout=30)
+    if r.status_code == 401:
+        erp_auth.login()
+        r = requests.get(url, headers=erp_auth.headers(), params=params or {}, timeout=30)
+    r.raise_for_status()
+    return r.json()
+
+
+def erp_post(path: str, data: dict):
+    """POST ERP 接口"""
+    url = f"{ERP_BASE_URL}{path}"
+    r = requests.post(url, headers={**erp_auth.headers(), "Content-Type": "application/json"},
+                      json=data, timeout=30)
+    if r.status_code == 401:
+        erp_auth.login()
+        r = requests.post(url, headers={**erp_auth.headers(), "Content-Type": "application/json"},
+                          json=data, timeout=30)
+    r.raise_for_status()
+    return r.json()
+
+
+# ─── 字典缓存 ─────────────────────────────────────────────────────────────────
+class DictCache:
+    """小字典分组的内存缓存，启动时加载，每日刷新"""
+
+    def __init__(self):
+        self.cache = {}          # {group: [name1, name2, ...]}
+        self.processes = []      # [{id, name, description}]  关联流程
+        self.last_loaded = 0
+        self.ttl = 24 * 3600     # 24 小时
+        self._lock = threading.Lock()
+
+    def load(self):
+        with self._lock:
+            logger.info("加载字典缓存...")
+            new_cache = {}
+            for group in SMALL_DICT_GROUPS:
+                try:
+                    data = erp_get("/api/backend/quick-inputs/", {"group": group, "limit": 200})
+                    items = data.get("results", [])
+                    new_cache[group] = [it["name"] for it in items if it.get("name")]
+                    logger.info(f"  [{group}] {len(new_cache[group])} 项")
+                except Exception as e:
+                    logger.warning(f"  [{group}] 加载失败: {e}")
+                    new_cache[group] = self.cache.get(group, [])
+            self.cache = new_cache
+
+            # 加载关联流程
+            try:
+                data = erp_get("/api/v1/stateflow/processes/", {"limit": 100})
+                self.processes = [
+                    {"id": p["id"], "name": p["name"], "description": p.get("description", "")}
+                    for p in data.get("results", [])
+                ]
+                logger.info(f"  [关联流程] {len(self.processes)} 项")
+            except Exception as e:
+                logger.warning(f"  [关联流程] 加载失败: {e}")
+
+            self.last_loaded = time.time()
+            logger.info("字典缓存加载完成")
+
+    def get(self, group: str):
+        if time.time() - self.last_loaded > self.ttl:
+            try:
+                self.load()
+            except Exception:
+                pass
+        return self.cache.get(group, [])
+
+    def get_processes(self):
+        return self.processes
+
+    def find_process_id(self, name: str):
+        """根据名称查找流程 ID（精确 + 模糊）"""
+        if not name:
+            return None
+        for p in self.processes:
+            if p["name"] == name:
+                return p["id"]
+        # 模糊匹配
+        for p in self.processes:
+            if name in p["name"] or p["name"] in name:
+                return p["id"]
+        return None
+
+    def all(self):
+        return dict(self.cache)
+
+
+dict_cache = DictCache()
+
+
+def background_init():
+    """后台线程：登录 ERP + 加载字典缓存（不阻塞 Flask 启动）"""
+    try:
+        erp_auth.login()
+        dict_cache.load()
+    except Exception as e:
+        logger.error(f"初始化失败: {e}")
+
+
+# ─── 布料/客户 智能匹配 ───────────────────────────────────────────────────────
+
+def find_or_create_fabric(name: str) -> dict:
+    """
+    精确匹配优先；找不到则新增（不做模糊匹配，避免错配）
+    返回 {id, name, exists, created}
+    """
+    if not name:
+        return None
+    name = name.strip()
+    try:
+        # 精确搜索（按 name 参数过滤）
+        data = erp_get("/api/backend/quick-inputs/", {"group": "布料", "name": name, "limit": 20})
+        items = data.get("results", [])
+        # 只接受精确匹配
+        exact = next((it for it in items if it.get("name") == name), None)
+        if exact:
+            return {"id": exact["id"], "name": exact["name"], "exists": True, "created": False}
+        # 找不到精确 → 新增
+        new_item = erp_post("/api/backend/quick-inputs/", {
+            "group": "布料",
+            "name":  name,
+            "value": name,
+        })
+        logger.info(f"新增布料: {name} (id={new_item.get('id')})")
+        return {"id": new_item.get("id"), "name": name, "exists": False, "created": True}
+    except Exception as e:
+        logger.error(f"匹配/新增布料失败 [{name}]: {e}")
+        return None
+
+
+def find_customer(name: str) -> dict:
+    """
+    搜索客户：精确匹配优先；模糊匹配只在精确未命中时使用，并标记 fuzzy
+    返回 {id, name, exists, ...}
+    """
+    if not name:
+        return None
+    name = name.strip()
+    try:
+        data = erp_get("/api/backend/customers/", {"name": name, "limit": 10})
+        items = data.get("results", []) if isinstance(data, dict) else data
+        if not items:
+            return {"id": None, "name": name, "exists": False}
+        # 精确匹配
+        exact = next((it for it in items if it.get("name") == name), None)
+        if exact:
+            return {
+                "id": exact["id"], "name": exact["name"], "exists": True,
+                "area": exact.get("area"),
+                "visible_employees": exact.get("visible_employees"),
+            }
+        # 没有精确匹配 → 返回未找到（让前端决定是否模糊建议）
+        return {
+            "id": None, "name": name, "exists": False,
+            "suggestions": [{"id": it["id"], "name": it["name"]} for it in items[:5]],
+        }
+    except Exception as e:
+        logger.error(f"搜索客户失败 [{name}]: {e}")
+        return None
+
+
+# ─── AI 解析（带枚举约束）────────────────────────────────────────────────────
+
+def build_system_prompt() -> str:
+    """根据缓存的字典动态构建 prompt，让 AI 严格选枚举"""
+
+    def opts(group):
+        vals = dict_cache.get(group)
+        return "/".join(vals) if vals else "（待加载）"
+
+    # 流程选项（带描述）
+    process_lines = []
+    for p in dict_cache.get_processes():
+        desc = f"({p['description']})" if p.get("description") else ""
+        process_lines.append(f"  - {p['name']}{desc}")
+    process_text = "\n".join(process_lines) if process_lines else "（待加载）"
+
+    return f"""你是一个印染厂开版订单信息提取助手。
+
+用户会发给你一条企业微信群聊消息，消息内容是跟单员发给设计师的工单指令。
+你需要从消息中提取以下字段，返回 JSON 格式。
+
+【字段说明（**必须严格使用以下枚举值**，不在列表中的返回 null）】
+
+- customer_name: 客户名称（自由文本，消息开头通常是客户名，如"華鑫"、"林庆福"、"糖糖"）
+- plate_type: 起版情况，**只能选**：{opts('起版情况')}
+    映射规则：图片起版/样衣起版/图片开版/样衣开版/找图开发/起版 → 首版
+              修改/加色/调色/调整/改版 → 修改单
+              套大货/套纸样/套样/重新套 → 复版/开版
+- urgency_level: 紧急程度，**只能选**：{opts('紧急程度')}
+    映射规则：消息有"加急"→ 加急；有"特急"→ 特急；"加急已下单"等 → 紧急已下单；否则 → 正常
+- production_method: 做货方式，**只能选**：{opts('做货方式')}
+- plate_method: 开版方式，**只能选**：{opts('开版方式')}
+    映射规则：消息有"图片起版/图片开版"→ 图片开版
+              消息有"样衣起版/样板起版/样衣开版"→ 样衣开版
+              消息有"文件开版"→ 文件开版
+              其他原始词如"开发"、"开板"→ 开发
+- area: 客户区域，**只能选**：{opts('地区')}（消息里很少提及，多数返回 null）
+- fabric: 面料名称（自由文本，完整保留克重和布料类型，如"140g四面弹"、"120克本白四面弹"、"180克牛奶丝"、"泡泡皱"）
+- width: 面料幅宽，**只能选**：{opts('幅宽')}（如"幅宽151"匹配"1.51"，"120"匹配"120"，找不到则 null）
+- fabric_source: 面料来源，**只能选**：{opts('布源')}
+- drawing_rating: 画图评级，**只能选**：{opts('画图评级')}（消息里几乎不会有，多返 null）
+- color_matching_rating: 调色评级，**只能选**：{opts('调色评级')}（多返 null）
+- sample_rating: 套样评级，**只能选**：{opts('套样评级')}（多返 null）
+- difficulty_rating: 难度评级，**只能选**：{opts('难度评级')}（多返 null）
+- process_name: 关联流程，**只能选**以下名称之一（直接返回流程名字符串，不要带括号注释）：
+{process_text}
+    选择规则：根据 plate_type 和操作内容综合判断
+              - 套大货/套纸样 → "开版(套米样)" 或 "开版(套大货)"
+              - 加色/调色 → "开版(配色)" 或 "开版(绘图调色)"
+              - P图/抠图 → "开版(P图)"
+              - 描图/找图开发 → "开版(描图开版)" 或 "开版(绘图调色套样)"
+              - 不确定时返回 null（让用户手动选）
+- plate_notes: 打版备注（自由文本，把消息中的具体要求、注意事项原文保留：克重、颜色要求、特殊工艺等）
+- style_name: 款号名称（如"NSY1090"、"G67单600-20款"、"6576改"等款号，没有则 null）
+- original_design_code: 原订单ID（如果消息提到要"套大货/套纸样/修改"等，且消息中有 5-7 位的纯数字编号，那就是原订单号；否则 null）
+- merchandiser_name: 跟单员（直接返回 sender 字段的原值）
+
+【重要规则】
+1. 枚举字段（plate_type、urgency_level、production_method、plate_method、area、width、fabric_source、各评级、process_name）**必须**从给定选项里选，不能自创；判断不出就返回 null
+2. 自由文本字段（customer_name、fabric、plate_notes、style_name、original_design_code）保持原文
+3. 一律返回 JSON 对象，不要任何其它文字"""
+
+
+def call_doubao(user_text: str) -> dict:
+    """调用豆包 /v3/responses，返回解析后的 dict"""
+    payload = {
+        "model": AI_MODEL,
+        "input": [
+            {"role": "system", "content": [{"type": "input_text", "text": build_system_prompt()}]},
+            {"role": "user",   "content": [{"type": "input_text", "text": user_text}]},
+        ],
+        "temperature": 0.1,
+        "max_output_tokens": 4096,
+    }
+    r = requests.post(DOUBAO_URL, headers=DOUBAO_HEADERS, json=payload, timeout=60)
+    r.raise_for_status()
+    data = r.json()
+
+    raw_text = ""
+    for item in data.get("output", []):
+        if item.get("type") == "message":
+            for c in item.get("content", []):
+                if c.get("type") == "output_text":
+                    raw_text = c.get("text", "").strip()
+                    break
+            break
+    if not raw_text:
+        raise ValueError("豆包返回内容为空")
+
+    start = raw_text.find("{")
+    end   = raw_text.rfind("}") + 1
+    if start == -1 or end == 0:
+        raise ValueError(f"响应中未找到 JSON：{raw_text}")
+    return json.loads(raw_text[start:end])
+
+
+# ─── 路由 ─────────────────────────────────────────────────────────────────────
+
+@app.route("/health", methods=["GET"])
+def health():
+    return jsonify({
+        "status": "ok",
+        "model": AI_MODEL,
+        "dict_cache_loaded": bool(dict_cache.cache),
+        "dict_groups": list(dict_cache.cache.keys()),
+        "erp_logged_in": bool(erp_auth.access_token),
+    })
+
+
+@app.route("/api/v1/dict-cache/", methods=["GET"])
+def get_dict_cache():
+    """查看当前缓存（调试用）"""
+    return jsonify(dict_cache.all())
+
+
+@app.route("/api/v1/dict-cache/refresh/", methods=["POST"])
+def refresh_dict_cache():
+    """手动刷新字典缓存"""
+    try:
+        dict_cache.load()
+        return jsonify({"status": "ok", "groups": list(dict_cache.cache.keys())})
+    except Exception as e:
+        return jsonify({"error": str(e)}), 500
+
+
+# ─── 企微消息代理 ─────────────────────────────────────────────────────────────
+
+@app.route("/api/v1/wechat/messages/", methods=["GET"])
+def proxy_wechat_messages():
+    limit      = request.args.get("limit", "30")
+    offset     = request.args.get("offset", "0")
+    room_id    = request.args.get("room_id", WORK_ORDER_ROOM)
+    with_image = request.args.get("with_image", "false")
+
+    try:
+        r = requests.get(
+            f"{WECHAT_API_BASE}/messages",
+            headers={"Authorization": WECHAT_AUTH},
+            params={"room_id": room_id, "limit": limit, "offset": offset, "with_image": with_image},
+            timeout=60,
+        )
+        r.raise_for_status()
+        return jsonify(r.json())
+    except requests.Timeout:
+        return jsonify({"error": "企微接口超时"}), 504
+    except Exception as e:
+        logger.error(f"代理请求失败: {e}")
+        return jsonify({"error": str(e)}), 500
+
+
+@app.route("/api/v1/wechat/messages/<path:msg_id>/with-image", methods=["GET"])
+def proxy_get_message_with_image(msg_id):
+    """
+    代理 GET /messages/:msg_id?with_image=true
+    企微服务直接支持按 msg_id 查单条消息，速度极快（< 0.1s）
+    """
+    try:
+        r = requests.get(
+            f"{WECHAT_API_BASE}/messages/{msg_id}",
+            headers={"Authorization": WECHAT_AUTH},
+            params={"with_image": "true"},
+            timeout=30,
+        )
+        r.raise_for_status()
+        return jsonify(r.json())
+    except requests.Timeout:
+        return jsonify({"error": "企微接口超时"}), 504
+    except requests.HTTPError as e:
+        return jsonify({"error": f"企微接口错误 {e.response.status_code}"}), e.response.status_code
+    except Exception as e:
+        logger.error(f"获取消息图片失败 msg_id={msg_id}: {e}")
+        return jsonify({"error": str(e)}), 500
+
+
+@app.route("/api/v1/wechat/messages/<path:msg_id>/tags", methods=["PUT"])
+def proxy_update_tags(msg_id):
+    try:
+        body = request.get_data()
+        r = requests.put(
+            f"{WECHAT_API_BASE}/messages/{msg_id}/tags",
+            headers={"Authorization": WECHAT_AUTH, "Content-Type": "application/json"},
+            data=body,
+            timeout=10,
+        )
+        return jsonify(r.json()), r.status_code
+    except Exception as e:
+        return jsonify({"error": str(e)}), 500
+
+
+@app.route("/api/v1/wechat/messages/<path:msg_id>/parse-record/", methods=["POST"])
+def record_parse(msg_id):
+    try:
+        body = request.get_json(silent=True) or {}
+        old_tags = body.get("old_tags") or {}
+        new_count = int(old_tags.get("parse_count", 0)) + 1
+        new_tags = {
+            **old_tags,
+            "parse_count": new_count,
+            "last_parsed_at": datetime.utcnow().isoformat() + "Z",
+            "last_parsed_by": body.get("user", "anonymous"),
+        }
+        r = requests.put(
+            f"{WECHAT_API_BASE}/messages/{msg_id}/tags",
+            headers={"Authorization": WECHAT_AUTH, "Content-Type": "application/json"},
+            json=new_tags,
+            timeout=10,
+        )
+        r.raise_for_status()
+        return jsonify(r.json())
+    except Exception as e:
+        logger.error(f"记录解析失败 msg_id={msg_id}: {e}")
+        return jsonify({"error": str(e)}), 500
+
+
+# ─── 主接口：解析开版订单 ────────────────────────────────────────────────────
+
+@app.route("/api/v1/ai/parse-plate-order/", methods=["POST"])
+def parse_plate_order():
+    """
+    解析企业微信消息 → 开版订单字段
+    自动匹配/新增布料、搜索客户
+
+    请求体：
+    {
+        "text": "消息文字",
+        "sender": "发送者",
+        "has_images": true,
+        "image_count": 2
+    }
+    """
+    data        = request.get_json(silent=True) or {}
+    text        = data.get("text", "").strip()
+    sender      = data.get("sender", "")
+    has_images  = data.get("has_images", False)
+    image_count = data.get("image_count", 0)
+
+    if not text and not has_images:
+        return jsonify({"error": "消息内容为空"}), 400
+
+    user_content = f"发送者：{sender}\n"
+    if has_images:
+        user_content += f"（消息含 {image_count} 张图片）\n"
+    user_content += f"消息内容：{text or '（纯图片消息，无文字）'}"
+
+    logger.info(f"解析请求 | sender={sender} | text={text[:80]}")
+
+    try:
+        result = call_doubao(user_content)
+    except Exception as e:
+        logger.error(f"AI 解析失败: {e}")
+        return jsonify({"error": f"AI 解析失败: {e}"}), 500
+
+    # 清理空值
+    cleaned = {k: v for k, v in result.items() if v not in (None, "", "null")}
+
+    # 跟单员兜底
+    if "merchandiser_name" not in cleaned and sender:
+        cleaned["merchandiser_name"] = sender
+
+    # 校验枚举字段（不在缓存里就剔除）
+    enum_fields = {
+        "plate_type": "起版情况",
+        "urgency_level": "紧急程度",
+        "production_method": "做货方式",
+        "plate_method": "开版方式",
+        "area": "地区",
+        "width": "幅宽",
+        "fabric_source": "布源",
+        "drawing_rating": "画图评级",
+        "color_matching_rating": "调色评级",
+        "sample_rating": "套样评级",
+        "difficulty_rating": "难度评级",
+    }
+    for field, group in enum_fields.items():
+        if field in cleaned:
+            valid = dict_cache.get(group)
+            if valid and cleaned[field] not in valid:
+                logger.warning(f"AI 返回 {field}={cleaned[field]} 不在枚举中，剔除")
+                cleaned.pop(field, None)
+
+    # 处理布料：搜索 → 不存在则新增
+    if cleaned.get("fabric"):
+        match = find_or_create_fabric(cleaned["fabric"])
+        if match:
+            cleaned["fabric"] = match["name"]   # 用规范化的名称
+            cleaned["fabric_id"] = match["id"]
+            cleaned["fabric_created"] = match.get("created", False)
+
+    # 处理客户：精确匹配优先；找不到时附带模糊建议供前端选择
+    if cleaned.get("customer_name"):
+        match = find_customer(cleaned["customer_name"])
+        if match and match.get("id"):
+            cleaned["customer_id"] = match["id"]
+            cleaned["customer_name"] = match["name"]
+            if match.get("area"):
+                cleaned["customer_area"] = match["area"]
+            if match.get("visible_employees"):
+                cleaned["customer_visible_employees"] = match["visible_employees"]
+        else:
+            cleaned["customer_not_found"] = True
+            if match and match.get("suggestions"):
+                cleaned["customer_suggestions"] = match["suggestions"]
+
+    # 处理关联流程：根据 process_name 查找 ID
+    if cleaned.get("process_name"):
+        pid = dict_cache.find_process_id(cleaned["process_name"])
+        if pid:
+            cleaned["process_id"] = pid
+        else:
+            logger.warning(f"未找到流程：{cleaned['process_name']}")
+            cleaned.pop("process_name", None)
+
+    logger.info(f"解析结果：{cleaned}")
+    return jsonify(cleaned)
+
+
+# ─── 入口 ─────────────────────────────────────────────────────────────────────
+
+if __name__ == "__main__":
+    port  = int(os.getenv("PORT", 5050))
+    debug = os.getenv("DEBUG", "false").lower() == "true"
+
+    # 后台线程初始化（不阻塞 Flask 启动）
+    threading.Thread(target=background_init, daemon=True).start()
+
+    logger.info(f"启动 plate-ai-service | port={port} | model={AI_MODEL}")
+    app.run(host="0.0.0.0", port=port, debug=debug, use_reloader=False)
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000..3fe7025
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,4 @@
+flask==3.0.3
+flask-cors==4.0.1
+requests==2.32.3
+python-dotenv==1.0.1
diff --git a/start.bat b/start.bat
new file mode 100644
index 0000000..12e9782
--- /dev/null
+++ b/start.bat
@@ -0,0 +1,8 @@
+@echo off
+cd /d %~dp0
+echo 安装依赖...
+pip install -r requirements.txt
+echo.
+echo 启动 plate-ai-service...
+python app.py
+pause