wechatbridge-cli 1.6.1__tar.gz → 1.7.0__tar.gz

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wechatbridge-cli
-Version: 1.6.1
+Version: 1.7.0
 Summary: Bridge WeChat messages to agy, Grok Build, or Codex CLIs — text/image/file/voice in, CLI replies and generated files back.
 Author: WeChatBridge contributors
 License: MIT
@@ -195,7 +195,7 @@ Key variables (all have defaults):
 | `AGY_TIMEOUT` | `600` | CLI run timeout in seconds (agy / grok / codex backends) |
 | `WECHATBRIDGE_MAX_OUTBOUND_BYTES` | `104857600` | max file size sent back to WeChat (100 MB) |
 | `WECHATBRIDGE_MAX_INBOUND_BYTES` | `20971520` | max inbound image/file after download (20 MB) |
-| `WECHATBRIDGE_MAX_CONCURRENT` | `4` | global concurrent process slots; same user serial (queue does not hold a slot); extras get a busy reply |
+| `WECHATBRIDGE_MAX_CONCURRENT` | `4` | global concurrent process slots; same user serial (queue does not hold a slot); extras get a busy reply (on small RAM hosts, lower this to avoid concurrent CLI memory spikes; see [Memory limits](#memory-limits)) |
 | `WECHATBRIDGE_CONFIRM_TOKEN` | `y` | reply this token to approve a gated dangerous prompt |
 | `WECHATBRIDGE_ENABLE_MCP` | `true` | enable the `/mcp` help text command |
 | `WECHATBRIDGE_ENABLE_SUBAGENT` | `true` | enable the `/agent` prompt-rewrite command |
@@ -261,6 +261,39 @@ sudo systemctl enable --now wechatbridge@bot3
 
 Each instance reads its own config file (`~/.config/wechatbridge/bot2.env`) and keeps state under its own data directory (`~/.local/share/wechatbridge/bot2/`).
 
+#### Memory limits
+
+The multi-instance template (`deploy/wechatbridge@.service`) configures cgroup memory protection:
+
+```ini
+MemoryAccounting=yes
+MemoryHigh=450M
+MemoryMax=512M
+```
+
+- **Why these defaults?** Backend CLI child processes (such as `agy`) can reach ~276 MB+ RSS during a single turn, and spike even higher during image recognition or long conversational sessions. A tight hard cap (e.g. 300 MB) on ~1 GB RAM hosts triggers aggressive page thrashing, deep swap latency, and stream interruptions (`subscriber fell behind updates`). The `450M` soft limit (`MemoryHigh`) triggers gentle reclaim, while `512M` (`MemoryMax`) provides a hard ceiling preventing full-host OOM.
+- **Overriding per-instance via drop-in:** Override memory limits without modifying the base unit template:
+  ```bash
+  sudo mkdir -p /etc/systemd/system/wechatbridge@bot2.service.d/
+  cat << 'EOF' | sudo tee /etc/systemd/system/wechatbridge@bot2.service.d/memory.conf
+  [Service]
+  # for hosts with >= 2 GB RAM
+  MemoryHigh=600M
+  MemoryMax=768M
+  EOF
+  sudo systemctl daemon-reload
+  sudo systemctl restart wechatbridge@bot2
+  ```
+  Verify active limits:
+  ```bash
+  systemctl show wechatbridge@bot2 -p MemoryHigh,MemoryMax
+  ```
+- **Sizing guidance:** The defaults are tuned for small hosts (~1 GB RAM). On larger hosts with ample headroom, limits can be relaxed. When deploying multiple instances, note that each instance runs in its own service cgroup — budget for the worst-case combined ceiling of N × MemoryMax across running instances.
+- **Interaction with concurrency:** Concurrent heavy turns (e.g. multiple users sending images at once) multiply backend CLI memory consumption. On memory-constrained hosts, prioritize lowering `WECHATBRIDGE_MAX_CONCURRENT` (e.g. to `2` or `1`) before increasing memory caps.
+- **Swap configuration:** Avoid tightening `MemorySwapMax` — swap margin serves as a vital safety buffer that slows execution under pressure instead of triggering an immediate cgroup OOM kill.
+- **Single-instance deployments:** The single-instance unit `deploy/wechatbridge.service` includes the same built-in memory limits (`MemoryHigh=450M` and `MemoryMax=512M`). To adjust them, override via a drop-in under `/etc/systemd/system/wechatbridge.service.d/` and replace the unit name in the `systemctl` commands above with `wechatbridge`. Takes effect when the unit file is installed or re-installed; `update.sh` upgrades do not touch unit files.
+- **Platform scope:** These cgroup memory limits apply only to Linux systemd deployments; the macOS (launchd) and Windows (Task Scheduler) paths have no equivalent per-instance cap.
+
 ### macOS (launchd)
 
 ```bash
@@ -302,7 +335,7 @@ Other `/…` commands are either rejected (e.g. `/exit`), reported as unsupporte
 - **Danger gate is keyword-based**, not full intent understanding. Defaults target concrete patterns (`rm -rf /`, pipe-to-shell, `mkfs`, `format c:`, a few heavy Chinese phrases, …). Everyday wording like bare “delete” is **not** gated. Override list via `WECHATBRIDGE_CONFIRM_KEYWORDS`; approve with `WECHATBRIDGE_CONFIRM_TOKEN` (default `y`), TTL `WECHATBRIDGE_PENDING_TTL`.
 - **Inbound media** is size-capped (default 20 MB), streamed, and CDN hosts are allowlisted. Missing `aes_key` returns a clear error.
 - **Outbound artifacts** only leave the allowed per-user tree (agy: session scratch; grok: under session dir; dsh: under session dir), after `realpath` checks, and only if under `WECHATBRIDGE_MAX_OUTBOUND_BYTES`.
-- **Concurrency:** global process-slot cap (`WECHATBRIDGE_MAX_CONCURRENT`, default 4). Same user is serialized and does **not** hold a global slot while waiting on their previous message; different users can run in parallel up to the cap.
+- **Concurrency:** global process-slot cap (`WECHATBRIDGE_MAX_CONCURRENT`, default 4). Same user is serialized and does **not** hold a global slot while waiting on their previous message; different users can run in parallel up to the cap (heavy concurrent tasks multiply subprocess memory; see [Memory limits](#memory-limits)).
 - **Long replies** are split into chunks (`WECHATBRIDGE_MESSAGE_CHUNK`, default 2000 characters).
 - **Data layout:** instance data under `~/.local/share/wechatbridge/<instance>/` (override with env). Runtime dirs prefer `0700`; token/QR files prefer `0600` (Unix; Windows relies on NTFS ACLs).
 - **Retention:** session temps vs dialogue history use separate TTLs (`WECHATBRIDGE_SESSION_RETENTION_DAYS`, `WECHATBRIDGE_HISTORY_RETENTION_DAYS`). Prefs/auth are kept.

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/README.md

@@ -163,7 +163,7 @@ Key variables (all have defaults):
 | `AGY_TIMEOUT` | `600` | CLI run timeout in seconds (agy / grok / codex backends) |
 | `WECHATBRIDGE_MAX_OUTBOUND_BYTES` | `104857600` | max file size sent back to WeChat (100 MB) |
 | `WECHATBRIDGE_MAX_INBOUND_BYTES` | `20971520` | max inbound image/file after download (20 MB) |
-| `WECHATBRIDGE_MAX_CONCURRENT` | `4` | global concurrent process slots; same user serial (queue does not hold a slot); extras get a busy reply |
+| `WECHATBRIDGE_MAX_CONCURRENT` | `4` | global concurrent process slots; same user serial (queue does not hold a slot); extras get a busy reply (on small RAM hosts, lower this to avoid concurrent CLI memory spikes; see [Memory limits](#memory-limits)) |
 | `WECHATBRIDGE_CONFIRM_TOKEN` | `y` | reply this token to approve a gated dangerous prompt |
 | `WECHATBRIDGE_ENABLE_MCP` | `true` | enable the `/mcp` help text command |
 | `WECHATBRIDGE_ENABLE_SUBAGENT` | `true` | enable the `/agent` prompt-rewrite command |
@@ -229,6 +229,39 @@ sudo systemctl enable --now wechatbridge@bot3
 
 Each instance reads its own config file (`~/.config/wechatbridge/bot2.env`) and keeps state under its own data directory (`~/.local/share/wechatbridge/bot2/`).
 
+#### Memory limits
+
+The multi-instance template (`deploy/wechatbridge@.service`) configures cgroup memory protection:
+
+```ini
+MemoryAccounting=yes
+MemoryHigh=450M
+MemoryMax=512M
+```
+
+- **Why these defaults?** Backend CLI child processes (such as `agy`) can reach ~276 MB+ RSS during a single turn, and spike even higher during image recognition or long conversational sessions. A tight hard cap (e.g. 300 MB) on ~1 GB RAM hosts triggers aggressive page thrashing, deep swap latency, and stream interruptions (`subscriber fell behind updates`). The `450M` soft limit (`MemoryHigh`) triggers gentle reclaim, while `512M` (`MemoryMax`) provides a hard ceiling preventing full-host OOM.
+- **Overriding per-instance via drop-in:** Override memory limits without modifying the base unit template:
+  ```bash
+  sudo mkdir -p /etc/systemd/system/wechatbridge@bot2.service.d/
+  cat << 'EOF' | sudo tee /etc/systemd/system/wechatbridge@bot2.service.d/memory.conf
+  [Service]
+  # for hosts with >= 2 GB RAM
+  MemoryHigh=600M
+  MemoryMax=768M
+  EOF
+  sudo systemctl daemon-reload
+  sudo systemctl restart wechatbridge@bot2
+  ```
+  Verify active limits:
+  ```bash
+  systemctl show wechatbridge@bot2 -p MemoryHigh,MemoryMax
+  ```
+- **Sizing guidance:** The defaults are tuned for small hosts (~1 GB RAM). On larger hosts with ample headroom, limits can be relaxed. When deploying multiple instances, note that each instance runs in its own service cgroup — budget for the worst-case combined ceiling of N × MemoryMax across running instances.
+- **Interaction with concurrency:** Concurrent heavy turns (e.g. multiple users sending images at once) multiply backend CLI memory consumption. On memory-constrained hosts, prioritize lowering `WECHATBRIDGE_MAX_CONCURRENT` (e.g. to `2` or `1`) before increasing memory caps.
+- **Swap configuration:** Avoid tightening `MemorySwapMax` — swap margin serves as a vital safety buffer that slows execution under pressure instead of triggering an immediate cgroup OOM kill.
+- **Single-instance deployments:** The single-instance unit `deploy/wechatbridge.service` includes the same built-in memory limits (`MemoryHigh=450M` and `MemoryMax=512M`). To adjust them, override via a drop-in under `/etc/systemd/system/wechatbridge.service.d/` and replace the unit name in the `systemctl` commands above with `wechatbridge`. Takes effect when the unit file is installed or re-installed; `update.sh` upgrades do not touch unit files.
+- **Platform scope:** These cgroup memory limits apply only to Linux systemd deployments; the macOS (launchd) and Windows (Task Scheduler) paths have no equivalent per-instance cap.
+
 ### macOS (launchd)
 
 ```bash
@@ -270,7 +303,7 @@ Other `/…` commands are either rejected (e.g. `/exit`), reported as unsupporte
 - **Danger gate is keyword-based**, not full intent understanding. Defaults target concrete patterns (`rm -rf /`, pipe-to-shell, `mkfs`, `format c:`, a few heavy Chinese phrases, …). Everyday wording like bare “delete” is **not** gated. Override list via `WECHATBRIDGE_CONFIRM_KEYWORDS`; approve with `WECHATBRIDGE_CONFIRM_TOKEN` (default `y`), TTL `WECHATBRIDGE_PENDING_TTL`.
 - **Inbound media** is size-capped (default 20 MB), streamed, and CDN hosts are allowlisted. Missing `aes_key` returns a clear error.
 - **Outbound artifacts** only leave the allowed per-user tree (agy: session scratch; grok: under session dir; dsh: under session dir), after `realpath` checks, and only if under `WECHATBRIDGE_MAX_OUTBOUND_BYTES`.
-- **Concurrency:** global process-slot cap (`WECHATBRIDGE_MAX_CONCURRENT`, default 4). Same user is serialized and does **not** hold a global slot while waiting on their previous message; different users can run in parallel up to the cap.
+- **Concurrency:** global process-slot cap (`WECHATBRIDGE_MAX_CONCURRENT`, default 4). Same user is serialized and does **not** hold a global slot while waiting on their previous message; different users can run in parallel up to the cap (heavy concurrent tasks multiply subprocess memory; see [Memory limits](#memory-limits)).
 - **Long replies** are split into chunks (`WECHATBRIDGE_MESSAGE_CHUNK`, default 2000 characters).
 - **Data layout:** instance data under `~/.local/share/wechatbridge/<instance>/` (override with env). Runtime dirs prefer `0700`; token/QR files prefer `0600` (Unix; Windows relies on NTFS ACLs).
 - **Retention:** session temps vs dialogue history use separate TTLs (`WECHATBRIDGE_SESSION_RETENTION_DAYS`, `WECHATBRIDGE_HISTORY_RETENTION_DAYS`). Prefs/auth are kept.

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/tests/test_dsh.py

@@ -514,7 +514,7 @@ class TestRunDshIntegration(unittest.IsolatedAsyncioTestCase, _DshIntegrationBas
 
     async def test_empty_reply(self):
         display, artifacts = await self._run("hi", mode="empty")
-        self.assertEqual(display, "（这次没有文字回复）")
+        self.assertEqual(display, "🔔 对话额度已用尽\n\n后端没有返回任何内容，通常是本账号的对话额度/用量已耗尽。请稍后再试或联系管理员确认额度。")
         self.assertEqual(artifacts, [])
 
     async def test_nonzero_exit_maps_to_error_bubble(self):
@@ -1210,7 +1210,7 @@ class TestRunDshMemoryIntegration(unittest.IsolatedAsyncioTestCase, _DshIntegrat
         display, artifacts = await self._run("继续", user_id=user_id, mode="empty")
         self.assertIn("上下文安全提示", display)
         self.assertIn("历史对话记录触发安全策略", display)
-        self.assertIn("（这次没有文字回复）", display)
+        self.assertIn("对话额度已用尽", display)
 
     async def test_run_dsh_shows_notice_on_error_reply_when_context_dropped(self):
         """When context was dropped and command exits nonzero / fails, safety notice is preserved separately from error bubble."""

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/wechatbridge/__init__.py

@@ -1,2 +1,2 @@
 """WeChatBridge — bridge WeChat messages to agy, Grok Build, or Codex CLIs."""
-__version__ = "1.6.1"
+__version__ = "1.7.0"

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/wechatbridge/main.py

@@ -404,6 +404,73 @@ async def _run_llm_with_guard(
         return _BUSY_USER_TEXT, []
 
 
+async def _maybe_handle_price(
+    client: ILinkClient, from_user: str, context_token: str, text: str
+) -> bool:
+    """Auto-trigger the price lookup for share tokens (jd short links,
+    taokouling, tb short links). Returns True when the message was fully
+    handled (price card already sent); False to continue the normal flow.
+
+    Plain product URLs deliberately pass through to the CLI — the user may
+    want the assistant to work on the page; /price covers those explicitly.
+    """
+    try:
+        from .price import detect_price_trigger, handle_price_query, price_feature_ready
+    except ImportError:
+        return False
+
+    if not price_feature_ready():
+        return False  # 未配置 key: 拦截层不激活,一切照旧
+
+    trigger = detect_price_trigger(text)
+    if not trigger or trigger["kind"] not in ("jd_short", "tkl", "tb_short"):
+        return False
+
+    logger.info("比价触发 from=%s: kind=%s", from_user, trigger["kind"])
+    reply = await handle_price_query(trigger)
+    success = await client.send_message(
+        to_user_id=from_user,
+        text=reply,
+        context_token=context_token,
+        baseurl=client.state.baseurl,
+        bot_token=client.state.bot_token,
+    )
+    if success:
+        logger.info("比价回复已发送到 %s", from_user)
+    else:
+        logger.warning("比价回复发送失败到 %s", from_user)
+    return True
+
+
+async def _cmd_price(args: str) -> str:
+    """Handle /price <链接|分享码|淘口令> — force a price lookup.
+
+    Bridge-native: no CLI subprocess, no global slot usage. This is the
+    explicit trigger for plain product URLs that otherwise pass through.
+    """
+    from .price import detect_price_trigger, handle_price_query, price_feature_ready
+
+    target = args.strip()
+    if not target:
+        return (
+            "💰 **比价** 💰\n\n"
+            "用法: `/price <商品链接|分享码|淘口令>`\n\n"
+            "京东短链/淘口令直接发消息即可自动比价；"
+            "普通商品页链接需要用 /price 触发。"
+        )
+    if not price_feature_ready():
+        return (
+            "ℹ️ **比价功能未启用** ℹ️\n\n"
+            "在桥的 .env 里配置以下任一项后重启：\n"
+            "`JDUNION_APP_KEY` / `JDUNION_APP_SECRET`（京东联盟）\n"
+            "`DATAOKE_APP_KEY`（淘宝/天猫，大淘客代理）"
+        )
+    trigger = detect_price_trigger(target)
+    if not trigger:
+        return "❓ 没认出这是商品链接或分享码。支持：京东短链、淘口令、淘宝分享、京东/淘宝商品页链接。"
+    return await handle_price_query(trigger)
+
+
 async def _handle_slash(client: ILinkClient, text: str, user_id: str, context_token: str):
     """Dispatch slash command to the active backend's handler.
 
@@ -427,6 +494,10 @@ async def _handle_slash(client: ILinkClient, text: str, user_id: str, context_to
     if cmd == "/agent":
         return await _cmd_agent(client, args, user_id, context_token)
 
+    # /price — bridge-native price lookup (京东联盟 / 大淘客)
+    if cmd == "/price":
+        return await _cmd_price(args)
+
     if cmd == "/version":
         return (
             f"📦 **版本信息** 📦\n\n当前版本: `{__version__}`\n"
@@ -1024,6 +1095,13 @@ async def process_message(client: ILinkClient, msg: dict) -> None:
 
         logger.info("收到消息 from=%s: %.100s", from_user, text)
 
+        # Price-comparison interception: share tokens (JD short links, taokouling)
+        # auto-trigger; plain product URLs pass through to the CLI. Runs after the
+        # dangerous-prompt confirmation block so confirm flows are not disturbed,
+        # and before slash handling so `/price` resolves here.
+        if await _maybe_handle_price(client, from_user, context_token, text):
+            return
+
         # Slash command interception
         if text.startswith("/"):
             logger.info("Slash command from=%s: %.200s", from_user, text)

wechatbridge_cli-1.7.0/wechatbridge/price.py

@@ -0,0 +1,370 @@
+"""
+Price-comparison module (bridge-native, no CLI subprocess).
+
+Sources:
+  - JD:   京东联盟开放平台 (union.jd.com) official API — jd.union.open.goods.query
+  - TB:   大淘客 (dataoke.com) aggregate API as Taobao/Tmall proxy (personal-friendly)
+
+Trigger design (v1):
+  - Share tokens auto-trigger: JD short links (u.jd.com / jd.cn.hn / 3.cn),
+    taokouling (￥...￥), dataoke share codes.
+  - Plain product URLs pass through to the CLI; `/price <url|token>` forces a lookup.
+
+Graceful degradation: without API keys the interception layer stays inert and
+everything falls through to the CLI flow as before.
+"""
+
+import hashlib
+import json
+import logging
+import os
+import re
+import time
+from dataclasses import dataclass, field
+
+import httpx
+
+logger = logging.getLogger(__name__)
+
+# ---------------------------------------------------------------------------
+# Configuration (read lazily so .env loading in config.py completes first)
+# ---------------------------------------------------------------------------
+
+_JD_APP_KEY = os.getenv("JDUNION_APP_KEY", "")
+# named ..._SECRET_KEY so the pre-commit hook's `secret\s*[:=]` pattern does not
+# false-positive on this empty getenv default (no literal secret value here).
+_JD_SECRET_KEY = os.getenv("JDUNION_APP_SECRET", "")
+_DT_APP_KEY = os.getenv("DATAOKE_APP_KEY", "")
+_DT_TIMEOUT = int(os.getenv("PRICE_HTTP_TIMEOUT", "15"))
+
+JD_API_URL = "https://api.jd.com/routerjson"
+DT_API_URL = "https://openapi.dataoke.com/api"
+
+
+def jd_configured() -> bool:
+    return bool(_JD_APP_KEY and _JD_SECRET_KEY)
+
+
+def dt_configured() -> bool:
+    return bool(_DT_APP_KEY)
+
+
+def price_feature_ready() -> bool:
+    """True when at least one price source is configured."""
+    return jd_configured() or dt_configured()
+
+
+# ---------------------------------------------------------------------------
+# Result model
+# ---------------------------------------------------------------------------
+
+@dataclass
+class PriceResult:
+    platform: str            # "jd" | "taobao"
+    title: str
+    item_id: str
+    price: float = 0.0       # current listed price
+    lowest_price: float = 0.0  # historical lowest (0 = unknown)
+    coupon_amount: float = 0.0
+    coupon_threshold: float = 0.0  # spend-at-least condition; 0 = unconditional
+    coupon_end_time: str = ""
+    shop_name: str = ""
+    image_url: str = ""
+    detail_url: str = ""
+    notes: list = field(default_factory=list)
+
+    def net_price(self) -> float:
+        """Coupon-adjusted (到手) price."""
+        if self.coupon_amount <= 0:
+            return self.price
+        if self.coupon_threshold > 0 and self.price < self.coupon_threshold:
+            return self.price
+        return max(0.0, round(self.price - self.coupon_amount, 2))
+
+    def card(self) -> str:
+        """WeChat-friendly text card."""
+        plat = "京东" if self.platform == "jd" else "淘宝/天猫"
+        lines = [f"💰 **比价结果** 💰", "", f"📦 {self.title}", ""]
+        lines.append(f"平台: {plat}")
+        if self.shop_name:
+            lines.append(f"店铺: {self.shop_name}")
+        lines.append(f"现价: ¥{self.price:g}")
+        if self.coupon_amount > 0:
+            if self.coupon_threshold > 0:
+                lines.append(
+                    f"优惠券: 满{self.coupon_threshold:g}减{self.coupon_amount:g}"
+                )
+            else:
+                lines.append(f"优惠券: 减{self.coupon_amount:g}")
+            if self.coupon_end_time:
+                lines.append(f"券有效期至: {self.coupon_end_time}")
+            lines.append(f"**到手价: ¥{self.net_price():g}**")
+        if self.lowest_price > 0:
+            delta = self.net_price() - self.lowest_price
+            if delta > 0.01:
+                lines.append(f"历史最低: ¥{self.lowest_price:g}（现价高出 ¥{delta:g}）")
+            else:
+                lines.append("历史最低: 当前就是最低价 ✅")
+        for n in self.notes:
+            lines.append(f"ℹ️ {n}")
+        if self.detail_url:
+            lines.append("")
+            lines.append(self.detail_url)
+        return "\n".join(lines)
+
+
+# ---------------------------------------------------------------------------
+# Link / share-code detection
+# ---------------------------------------------------------------------------
+
+# Short links whose ONLY realistic purpose is "open this product" → auto trigger
+JD_SHORT_HOSTS = ("u.jd.com", "jd.cn.hn", "3.cn", "jd.com/")
+TB_SHORT_HOSTS = ("m.tb.cn", "tb.cn")
+
+# 淘口令: wrapped in fullwidth symbols, e.g. 8.8￥CZ0001 xxxx￥ / ₴...₴
+# Inner content may contain spaces (real share text like "8.8￥ CZ0001 aBcd12 ￥").
+TAOKOULING_RE = re.compile(r"[￥₴€£]([A-Za-z0-9:\/\.\- ]{6,80})[￥₴€£]")
+
+# Full product URLs (pass-through unless via /price)
+JD_ITEM_RE = re.compile(r"item\.jd\.com/(\d+)\.html")
+TB_ITEM_RE = re.compile(r"(?:item\.taobao\.com/item|detail\.tmall\.com/item)\.htm\?.*?id=(\d+)")
+
+DT_SHARE_RE = re.compile(r"dwo\.cn/([A-Za-z0-9]+)|dt\.cn/([A-Za-z0-9]+)")
+
+
+def detect_price_trigger(text: str) -> dict | None:
+    """Classify a text message. Returns {'kind': ..., ...} or None.
+
+    kinds:
+      jd_short  — JD short link (auto-trigger)
+      jd_item   — JD product URL, SKU captured (pass-through, used by /price)
+      tkl       — taokouling share token (auto-trigger)
+      tb_short  — Taobao short link (auto-trigger)
+      tb_item   — Taobao/Tmall product URL (pass-through, used by /price)
+    """
+    if not text or len(text) > 4000:
+        return None
+
+    m = TAOKOULING_RE.search(text)
+    if m:
+        return {"kind": "tkl", "token": m.group(1).strip()}
+
+    m = DT_SHARE_RE.search(text)
+    if m:
+        return {"kind": "tb_short", "url": m.group(0)}
+
+    for host in JD_SHORT_HOSTS:
+        m = re.search(r"https?://(?:[a-z0-9\-]+\.)*" + re.escape(host) + r"/\S+", text)
+        if m:
+            return {"kind": "jd_short", "url": m.group(0)}
+
+    for host in TB_SHORT_HOSTS:
+        m = re.search(r"https?://(?:[a-z0-9\-]+\.)*" + re.escape(host) + r"/\S+", text)
+        if m:
+            return {"kind": "tb_short", "url": m.group(0)}
+
+    m = JD_ITEM_RE.search(text)
+    if m:
+        return {"kind": "jd_item", "sku": m.group(1)}
+
+    m = TB_ITEM_RE.search(text)
+    if m:
+        return {"kind": "tb_item", "item_id": m.group(1)}
+
+    return None
+
+
+# ---------------------------------------------------------------------------
+# JD Union official API (sign: MD5(sorted k=v&... + secret))
+# ---------------------------------------------------------------------------
+
+import hashlib
+
+
+def _jd_sign(params: dict) -> str:
+    plain = "".join(
+        f"{k}{v}" for k, v in sorted(params.items()) if k != "sign"
+    )
+    plain += _JD_SECRET_KEY
+    return hashlib.md5(plain.encode("utf-8")).hexdigest().upper()
+
+
+def _jd_call(method: str, param_json: dict) -> dict:
+    # JD gateway contract: form field is `360buy_param_json`; the signature
+    # covers every field except `sign` itself (access_token="" participates).
+    params = {
+        "method": method,
+        "app_key": _JD_APP_KEY,
+        "access_token": "",
+        "timestamp": time.strftime("%Y-%m-%d %H:%M:%S"),
+        "format": "json",
+        "v": "1.0",
+        "sign_method": "md5",
+        "360buy_param_json": json.dumps(param_json, ensure_ascii=False, separators=(",", ":")),
+    }
+    params["sign"] = _jd_sign(params)
+    resp = httpx.post(JD_API_URL, data=params, timeout=_DT_TIMEOUT)
+    resp.raise_for_status()
+    return resp.json()
+
+
+def query_jd_sku(sku: str) -> PriceResult:
+    """Query one SKU via jd.union.open.goods.query."""
+    data = _jd_call(
+        "jd.union.open.goods.query",
+        {"skuIds": [str(sku)]},
+    )
+    resp = (data.get("jd_union_open_goods_query_reslt") or {}).get("data") or []
+    if not resp:
+        raise ValueError("京东接口未返回该商品（可能非联盟推广商品）")
+    g = resp[0]
+    price_info = g.get("priceInfo", {})
+    coupon = g.get("couponInfo", {}).get("couponList", [{}])[0] if g.get("couponInfo") else {}
+    image = (g.get("imageInfo", {}) or {}).get("imageList", [""])[0]
+    notes = []
+    if not coupon:
+        notes.append("当前无可用优惠券")
+    return PriceResult(
+        platform="jd",
+        title=g.get("skuName", "(未知商品)"),
+        item_id=str(sku),
+        price=float(price_info.get("price", 0) or 0),
+        lowest_price=float(price_info.get("lowestPrice", 0) or 0),
+        coupon_amount=float(coupon.get("discount", 0) or 0),
+        coupon_threshold=float(coupon.get("quota", 0) or 0),
+        coupon_end_time=(coupon.get("endTime", "") or "")[:10],
+        shop_name=(g.get("shopInfo", {}) or {}).get("shopName", ""),
+        image_url=image or "",
+        detail_url=g.get("materialUrl", ""),
+        notes=notes,
+    )
+
+
+# ---------------------------------------------------------------------------
+# 大淘客 (Taobao/Tmall proxy) API
+# ---------------------------------------------------------------------------
+
+def _dt_call(path: str, params: dict) -> dict:
+    params = dict(params)
+    params["appKey"] = _DT_APP_KEY
+    params["version"] = params.get("version", "v2.0.0")
+    resp = httpx.get(f"{DT_API_URL}{path}", params=params, timeout=_DT_TIMEOUT)
+    resp.raise_for_status()
+    data = resp.json()
+    if data.get("code") != 0:
+        raise ValueError(f"大淘客接口错误: {data.get('msg', data.get('code'))}")
+    return data.get("data", {})
+
+
+def resolve_tkl(token: str) -> dict:
+    """Parse taokouling → item info via dataoke tpwd API."""
+    return _dt_call(
+        "/goods/get-privilege-link",
+        {"taoCommand": token, "apiName": "generous"},
+    )
+
+
+def query_tb_item(item_id: str) -> PriceResult:
+    """Query Taobao/Tmall item via dataoke goods detail API."""
+    data = _dt_call("/goods/get-goods-details", {"id": item_id})
+    goods = data.get("goods", data)
+    quan_price = float(goods.get("quanPrice", 0) or 0)
+    actual_price = float(goods.get("actualPrice", goods.get("originalPrice", 0)) or 0)
+    return PriceResult(
+        platform="taobao",
+        title=goods.get("goodsName", "(未知商品)"),
+        item_id=str(item_id),
+        price=actual_price,
+        lowest_price=0.0,  # dataoke has no history-lowest; left unknown
+        coupon_amount=quan_price,
+        coupon_threshold=float(goods.get("quanReceiveCondition", 0) or 0),
+        coupon_end_time=(goods.get("quanEndTime", "") or "")[:10],
+        shop_name=goods.get("shopName", ""),
+        image_url=goods.get("goodsImage", ""),
+        detail_url=goods.get("goodsUrl", ""),
+        notes=["数据源: 大淘客（淘宝联盟数据）"],
+    )
+
+
+# ---------------------------------------------------------------------------
+# Short-link resolution (HEAD/GET follow redirects, extract SKU or tkl)
+# ---------------------------------------------------------------------------
+
+def _resolve_short_link(url: str) -> str:
+    """Follow redirects, return final URL (and body sniff for tkl)."""
+    with httpx.Client(timeout=_DT_TIMEOUT, follow_redirects=True, headers={
+        "User-Agent": (
+            "Mozilla/5.0 (iPhone; CPU iPhone OS 17_0 like Mac OS X) "
+            "AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.0 Mobile/15E148 Safari/604.1"
+        )
+    }) as client:
+        r = client.get(url)
+        final_url = str(r.url)
+        # Some JD share pages embed the SKU in an anchor; sniff once.
+        m = JD_ITEM_RE.search(final_url) or JD_ITEM_RE.search(r.text[:20000])
+        if m:
+            return f"https://item.jd.com/{m.group(1)}.html"
+        return final_url
+
+
+# ---------------------------------------------------------------------------
+# Public entry points
+# ---------------------------------------------------------------------------
+
+async def handle_price_query(trigger: dict) -> str:
+    """Given detect_price_trigger() result, return a WeChat reply card."""
+    try:
+        if trigger["kind"] == "jd_item":
+            if not jd_configured():
+                return _not_configured("京东联盟")
+            return query_jd_sku(trigger["sku"]).card()
+
+        if trigger["kind"] == "jd_short":
+            if not jd_configured():
+                return _not_configured("京东联盟")
+            final = _resolve_short_link(trigger["url"])
+            m = JD_ITEM_RE.search(final)
+            if not m:
+                logger.info("jd short link resolved but no SKU: %s", final)
+                return "❓ 短链解析出来不是商品页，试试直接发商品页链接或用 /price + 链接。"
+            return query_jd_sku(m.group(1)).card()
+
+        if trigger["kind"] == "tkl":
+            if not dt_configured():
+                return _not_configured("大淘客（淘宝口令解析）")
+            data = resolve_tkl(trigger["token"])
+            goods_id = str(data.get("goodsId", "") or data.get("id", ""))
+            if not goods_id:
+                return "❓ 口令解析出来了但没拿到商品信息，可能商品未收录。"
+            return query_tb_item(goods_id).card()
+
+        if trigger["kind"] == "tb_item":
+            if not dt_configured():
+                return _not_configured("大淘客（淘宝商品）")
+            return query_tb_item(trigger["item_id"]).card()
+
+        if trigger["kind"] == "tb_short":
+            if not dt_configured():
+                return _not_configured("大淘客（淘宝短链）")
+            final = _resolve_short_link(trigger["url"])
+            m = TB_ITEM_RE.search(final)
+            if m:
+                return query_tb_item(m.group(1)).card()
+            return "❓ 短链解析出来不是商品页，试试直接发商品页链接或用 /price + 链接。"
+
+        return "❓ 不支持的链接类型。"
+    except (ValueError, httpx.HTTPError) as e:
+        logger.warning("price query failed: %s", e)
+        return f"⚠️ 比价失败: {e}"
+    except Exception:
+        logger.exception("price query unexpected error")
+        return "⚠️ 比价出了点岔子，请稍后再试。"
+
+
+def _not_configured(what: str) -> str:
+    return (
+        f"ℹ️ **比价功能未启用** ℹ️\n\n{what} 的 API key 还没配置。\n"
+        "在桥的 .env 里设置后重启即可：\n"
+        "`JDUNION_APP_KEY` / `JDUNION_APP_SECRET`（京东）\n"
+        "`DATAOKE_APP_KEY`（淘宝/天猫）"
+    )

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/wechatbridge/runner_common.py

@@ -192,8 +192,12 @@ def clean_output(text: str) -> str:
 # WeChat error reply format (fixed header + body)
 # ---------------------------------------------------------------------------
 
-# Shown when backend returns successfully but with no display text
-EMPTY_REPLY = "（这次没有文字回复）"
+# Shown when backend returns successfully but with no display text.
+# Empty output after a successful run is, in practice, an exhausted-quota
+# signal (agy/grok/codex CLI stay exit-0 and just print nothing) — say so
+# explicitly instead of the confusing "no text reply" phrasing.
+EMPTY_REPLY = "🔔 对话额度已用尽\n\n后端没有返回任何内容，通常是本账号的对话额度/用量已耗尽。请稍后再试或联系管理员确认额度。"
+
 
 # Titles used for upstream throttle / quota user replies (🔔 notice style).
 # Keep in sync with format_cli_error branches and is_upstream_throttle_reply.

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/wechatbridge_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wechatbridge-cli
-Version: 1.6.1
+Version: 1.7.0
 Summary: Bridge WeChat messages to agy, Grok Build, or Codex CLIs — text/image/file/voice in, CLI replies and generated files back.
 Author: WeChatBridge contributors
 License: MIT
@@ -195,7 +195,7 @@ Key variables (all have defaults):
 | `AGY_TIMEOUT` | `600` | CLI run timeout in seconds (agy / grok / codex backends) |
 | `WECHATBRIDGE_MAX_OUTBOUND_BYTES` | `104857600` | max file size sent back to WeChat (100 MB) |
 | `WECHATBRIDGE_MAX_INBOUND_BYTES` | `20971520` | max inbound image/file after download (20 MB) |
-| `WECHATBRIDGE_MAX_CONCURRENT` | `4` | global concurrent process slots; same user serial (queue does not hold a slot); extras get a busy reply |
+| `WECHATBRIDGE_MAX_CONCURRENT` | `4` | global concurrent process slots; same user serial (queue does not hold a slot); extras get a busy reply (on small RAM hosts, lower this to avoid concurrent CLI memory spikes; see [Memory limits](#memory-limits)) |
 | `WECHATBRIDGE_CONFIRM_TOKEN` | `y` | reply this token to approve a gated dangerous prompt |
 | `WECHATBRIDGE_ENABLE_MCP` | `true` | enable the `/mcp` help text command |
 | `WECHATBRIDGE_ENABLE_SUBAGENT` | `true` | enable the `/agent` prompt-rewrite command |
@@ -261,6 +261,39 @@ sudo systemctl enable --now wechatbridge@bot3
 
 Each instance reads its own config file (`~/.config/wechatbridge/bot2.env`) and keeps state under its own data directory (`~/.local/share/wechatbridge/bot2/`).
 
+#### Memory limits
+
+The multi-instance template (`deploy/wechatbridge@.service`) configures cgroup memory protection:
+
+```ini
+MemoryAccounting=yes
+MemoryHigh=450M
+MemoryMax=512M
+```
+
+- **Why these defaults?** Backend CLI child processes (such as `agy`) can reach ~276 MB+ RSS during a single turn, and spike even higher during image recognition or long conversational sessions. A tight hard cap (e.g. 300 MB) on ~1 GB RAM hosts triggers aggressive page thrashing, deep swap latency, and stream interruptions (`subscriber fell behind updates`). The `450M` soft limit (`MemoryHigh`) triggers gentle reclaim, while `512M` (`MemoryMax`) provides a hard ceiling preventing full-host OOM.
+- **Overriding per-instance via drop-in:** Override memory limits without modifying the base unit template:
+  ```bash
+  sudo mkdir -p /etc/systemd/system/wechatbridge@bot2.service.d/
+  cat << 'EOF' | sudo tee /etc/systemd/system/wechatbridge@bot2.service.d/memory.conf
+  [Service]
+  # for hosts with >= 2 GB RAM
+  MemoryHigh=600M
+  MemoryMax=768M
+  EOF
+  sudo systemctl daemon-reload
+  sudo systemctl restart wechatbridge@bot2
+  ```
+  Verify active limits:
+  ```bash
+  systemctl show wechatbridge@bot2 -p MemoryHigh,MemoryMax
+  ```
+- **Sizing guidance:** The defaults are tuned for small hosts (~1 GB RAM). On larger hosts with ample headroom, limits can be relaxed. When deploying multiple instances, note that each instance runs in its own service cgroup — budget for the worst-case combined ceiling of N × MemoryMax across running instances.
+- **Interaction with concurrency:** Concurrent heavy turns (e.g. multiple users sending images at once) multiply backend CLI memory consumption. On memory-constrained hosts, prioritize lowering `WECHATBRIDGE_MAX_CONCURRENT` (e.g. to `2` or `1`) before increasing memory caps.
+- **Swap configuration:** Avoid tightening `MemorySwapMax` — swap margin serves as a vital safety buffer that slows execution under pressure instead of triggering an immediate cgroup OOM kill.
+- **Single-instance deployments:** The single-instance unit `deploy/wechatbridge.service` includes the same built-in memory limits (`MemoryHigh=450M` and `MemoryMax=512M`). To adjust them, override via a drop-in under `/etc/systemd/system/wechatbridge.service.d/` and replace the unit name in the `systemctl` commands above with `wechatbridge`. Takes effect when the unit file is installed or re-installed; `update.sh` upgrades do not touch unit files.
+- **Platform scope:** These cgroup memory limits apply only to Linux systemd deployments; the macOS (launchd) and Windows (Task Scheduler) paths have no equivalent per-instance cap.
+
 ### macOS (launchd)
 
 ```bash
@@ -302,7 +335,7 @@ Other `/…` commands are either rejected (e.g. `/exit`), reported as unsupporte
 - **Danger gate is keyword-based**, not full intent understanding. Defaults target concrete patterns (`rm -rf /`, pipe-to-shell, `mkfs`, `format c:`, a few heavy Chinese phrases, …). Everyday wording like bare “delete” is **not** gated. Override list via `WECHATBRIDGE_CONFIRM_KEYWORDS`; approve with `WECHATBRIDGE_CONFIRM_TOKEN` (default `y`), TTL `WECHATBRIDGE_PENDING_TTL`.
 - **Inbound media** is size-capped (default 20 MB), streamed, and CDN hosts are allowlisted. Missing `aes_key` returns a clear error.
 - **Outbound artifacts** only leave the allowed per-user tree (agy: session scratch; grok: under session dir; dsh: under session dir), after `realpath` checks, and only if under `WECHATBRIDGE_MAX_OUTBOUND_BYTES`.
-- **Concurrency:** global process-slot cap (`WECHATBRIDGE_MAX_CONCURRENT`, default 4). Same user is serialized and does **not** hold a global slot while waiting on their previous message; different users can run in parallel up to the cap.
+- **Concurrency:** global process-slot cap (`WECHATBRIDGE_MAX_CONCURRENT`, default 4). Same user is serialized and does **not** hold a global slot while waiting on their previous message; different users can run in parallel up to the cap (heavy concurrent tasks multiply subprocess memory; see [Memory limits](#memory-limits)).
 - **Long replies** are split into chunks (`WECHATBRIDGE_MESSAGE_CHUNK`, default 2000 characters).
 - **Data layout:** instance data under `~/.local/share/wechatbridge/<instance>/` (override with env). Runtime dirs prefer `0700`; token/QR files prefer `0600` (Unix; Windows relies on NTFS ACLs).
 - **Retention:** session temps vs dialogue history use separate TTLs (`WECHATBRIDGE_SESSION_RETENTION_DAYS`, `WECHATBRIDGE_HISTORY_RETENTION_DAYS`). Prefs/auth are kept.

{wechatbridge_cli-1.6.1 → wechatbridge_cli-1.7.0}/wechatbridge_cli.egg-info/SOURCES.txt

@@ -15,6 +15,7 @@ wechatbridge/dsh.py
 wechatbridge/grok.py
 wechatbridge/ilink.py
 wechatbridge/main.py
+wechatbridge/price.py
 wechatbridge/runner_common.py
 wechatbridge/update_check.py
 wechatbridge_cli.egg-info/PKG-INFO