ff

app.py

@@ -6,14 +6,15 @@ import threading
 import httpx
 import logging
 import gradio as gr
+from fastapi import FastAPI, Request
+from fastapi.responses import PlainTextResponse
 from openai import OpenAI
 from Crypto.Cipher import AES
 from Crypto.Util.Padding import unpad
-from fastapi import Request
-from fastapi.responses import PlainTextResponse
 import xml.etree.ElementTree as ET
+import uvicorn
 
-# ── 日志配置 ──────────────────────────────────────────────────
+# ── 日志 ──────────────────────────────────────────────────────
 logging.basicConfig(
     level=logging.INFO,
     format="%(asctime)s [%(levelname)s] %(message)s",
@@ -29,11 +30,8 @@ WECOM_CORPID   = os.getenv("CORPID", "").strip()
 WECOM_SECRET   = os.getenv("WECOM_SECRET", "").strip()
 WECOM_AGENTID  = os.getenv("WECOM_AGENTID", "1000003")
 
-# ── LLM 客户端 ────────────────────────────────────────────────
-llm = OpenAI(
-    base_url="https://integrate.api.nvidia.com/v1",
-    api_key=NVIDIA_API_KEY,
-)
+# ── LLM ───────────────────────────────────────────────────────
+llm = OpenAI(base_url="https://integrate.api.nvidia.com/v1", api_key=NVIDIA_API_KEY)
 
 SYSTEM_PROMPT = (
     "You are Hermes, a helpful AI assistant. "
@@ -60,7 +58,7 @@ def chat_with_hermes(message: str, history: list) -> str:
     log.info("[WEB] assistant: %s", reply)
     return reply
 
-# ── 企业微信工具函数 ──────────────────────────────────────────
+# ── 企业微信 ──────────────────────────────────────────────────
 _wx_cache = {"token": "", "expires": 0}
 _lock = threading.Lock()
 
@@ -74,20 +72,20 @@ def get_wecom_token() -> str:
             params={"corpid": WECOM_CORPID, "corpsecret": WECOM_SECRET},
             timeout=10,
         ).json()
-        log.info("[WECOM] get_token response: %s", r)
+        log.info("[WECOM] get_token: %s", r)
         _wx_cache["token"] = r.get("access_token", "")
         _wx_cache["expires"] = time.time() + r.get("expires_in", 7200)
         return _wx_cache["token"]
 
 def send_wecom_message(to_user: str, content: str):
     token = get_wecom_token()
-    payload = {"touser": to_user, "msgtype": "text",
-               "agentid": int(WECOM_AGENTID), "text": {"content": content}}
     r = httpx.post(
         f"https://qyapi.weixin.qq.com/cgi-bin/message/send?access_token={token}",
-        json=payload, timeout=10,
+        json={"touser": to_user, "msgtype": "text",
+              "agentid": int(WECOM_AGENTID), "text": {"content": content}},
+        timeout=10,
     ).json()
-    log.info("[WECOM] send_message to=%s result=%s", to_user, r)
+    log.info("[WECOM] send to=%s result=%s", to_user, r)
 
 def decrypt_msg(encrypted: str) -> str:
     aes_key = base64.b64decode(WECOM_AES_KEY + "=")
@@ -100,42 +98,37 @@ def verify_sig(timestamp: str, nonce: str, encrypt: str = "") -> str:
     parts = [WECOM_TOKEN, timestamp, nonce] + ([encrypt] if encrypt else [])
     return hashlib.sha1("".join(sorted(parts)).encode()).hexdigest()
 
-# ── Gradio 界面 ───────────────────────────────────────────────
-demo = gr.ChatInterface(
-    fn=chat_with_hermes,
-    title="🤖 Hermes Agent",
-    description="Powered by NVIDIA NIM · 企业微信同步接入",
-)
+# ── FastAPI ───────────────────────────────────────────────────
+fastapi_app = FastAPI()
 
-# ── 挂载企业微信路由到 Gradio 内置 FastAPI app ────────────────
-@demo.app.get("/gateway/wecom")
+@fastapi_app.get("/gateway/wecom")
 async def wecom_verify(request: Request):
     p = request.query_params
-    log.info("[WECOM] GET verify params: %s", dict(p))
+    log.info("[WECOM] GET params: %s", dict(p))
     try:
-        sig = verify_sig(p.get("timestamp",""), p.get("nonce",""), p.get("echostr",""))
-        expected = p.get("msg_signature","")
-        log.info("[WECOM] verify sig=%s expected=%s match=%s", sig, expected, sig==expected)
+        sig = verify_sig(p.get("timestamp", ""), p.get("nonce", ""), p.get("echostr", ""))
+        expected = p.get("msg_signature", "")
+        log.info("[WECOM] sig=%s expected=%s match=%s", sig, expected, sig == expected)
         if sig != expected:
             return PlainTextResponse("")
-        plain = decrypt_msg(p.get("echostr",""))
-        log.info("[WECOM] verify ok, echostr plain=%s", plain)
+        plain = decrypt_msg(p.get("echostr", ""))
+        log.info("[WECOM] verify ok plain=%s", plain)
         return PlainTextResponse(plain)
     except Exception as e:
         log.exception("[WECOM] verify error: %s", e)
         return PlainTextResponse("")
 
-@demo.app.post("/gateway/wecom")
+@fastapi_app.post("/gateway/wecom")
 async def wecom_message(request: Request):
     p = request.query_params
     log.info("[WECOM] POST params: %s", dict(p))
     try:
         body = await request.body()
-        log.info("[WECOM] POST body: %s", body.decode())
+        log.info("[WECOM] body: %s", body.decode())
         encrypt = ET.fromstring(body.decode()).findtext("Encrypt", "")
-        sig = verify_sig(p.get("timestamp",""), p.get("nonce",""), encrypt)
-        expected = p.get("msg_signature","")
-        log.info("[WECOM] msg sig=%s expected=%s match=%s", sig, expected, sig==expected)
+        sig = verify_sig(p.get("timestamp", ""), p.get("nonce", ""), encrypt)
+        expected = p.get("msg_signature", "")
+        log.info("[WECOM] sig=%s expected=%s match=%s", sig, expected, sig == expected)
         if sig != expected:
             return PlainTextResponse("success")
         root = ET.fromstring(decrypt_msg(encrypt))
@@ -144,12 +137,12 @@ async def wecom_message(request: Request):
         log.info("[WECOM] msg_type=%s from=%s", msg_type, from_user)
         if msg_type == "text":
             content = root.findtext("Content", "").strip()
-            log.info("[WECOM] user=%s msg=%s", from_user, content)
+            log.info("[WECOM] content=%s", content)
             if content:
                 def reply():
                     try:
                         answer = chat_with_hermes(content, [])
-                        log.info("[WECOM] reply to=%s answer=%s", from_user, answer)
+                        log.info("[WECOM] reply=%s", answer)
                         send_wecom_message(from_user, answer)
                     except Exception as e:
                         log.exception("[WECOM] reply error: %s", e)
@@ -158,14 +151,17 @@ async def wecom_message(request: Request):
         log.exception("[WECOM] message error: %s", e)
     return PlainTextResponse("success")
 
+# ── Gradio 挂载到 FastAPI ─────────────────────────────────────
+demo = gr.ChatInterface(
+    fn=chat_with_hermes,
+    title="🤖 Hermes Agent",
+    description="Powered by NVIDIA NIM · 企业微信同步接入",
+)
+app = gr.mount_gradio_app(fastapi_app, demo, path="/")
+
 # ── 启动 ──────────────────────────────────────────────────────
 if __name__ == "__main__":
-    log.info("Starting Hermes Agent...")
-    log.info("WECOM_TOKEN set: %s", bool(WECOM_TOKEN))
-    log.info("WECOM_AES_KEY set: %s", bool(WECOM_AES_KEY))
-    log.info("WECOM_CORPID set: %s", bool(WECOM_CORPID))
-    log.info("WECOM_SECRET set: %s", bool(WECOM_SECRET))
-    log.info("NVIDIA_API_KEY set: %s", bool(NVIDIA_API_KEY))
-    routes = [r.path for r in demo.app.routes]
-    log.info("Registered routes: %s", routes)
-    demo.launch(server_name="0.0.0.0", server_port=7860)
+    log.info("WECOM_TOKEN=%s AES_KEY=%s CORPID=%s SECRET=%s NVIDIA=%s",
+             bool(WECOM_TOKEN), bool(WECOM_AES_KEY),
+             bool(WECOM_CORPID), bool(WECOM_SECRET), bool(NVIDIA_API_KEY))
+    uvicorn.run(app, host="0.0.0.0", port=7860)