From c1bf38a79b87ce7f2b680c2cde4fb3432c968287 Mon Sep 17 00:00:00 2001
From: "kyj@bowong.ai" <kyj@bowong.ai>
Date: Thu, 3 Apr 2025 18:04:42 +0800
Subject: [PATCH] =?UTF-8?q?ADD=20=E5=A2=9E=E5=8A=A0HeyGem=E5=8F=A3?=
 =?UTF-8?q?=E5=9E=8B=E5=90=8C=E6=AD=A5(API)=E8=8A=82=E7=82=B9?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 __init__.py     |  7 ++--
 nodes/heygem.py | 92 +++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 97 insertions(+), 2 deletions(-)
 create mode 100644 nodes/heygem.py

diff --git a/__init__.py b/__init__.py
index 17176ce..dabdb9a 100644
--- a/__init__.py
+++ b/__init__.py
@@ -1,3 +1,4 @@
+from .nodes.heygem import HeyGemF2F
 from .nodes.s3 import S3Download, S3Upload
 from .nodes.text import *
 from .nodes.traverse_folder import TraverseFolder
@@ -29,7 +30,8 @@ NODE_CLASS_MAPPINGS = {
     "unloadAllModels": UnloadAllModels,
     "TraverseFolder": TraverseFolder,
     "LoadTextCustom": LoadTextLocal,
-    "LoadTextCustomOnline": LoadTextOnline
+    "LoadTextCustomOnline": LoadTextOnline,
+    "HeyGemF2F": HeyGemF2F,
 }
 
 # A dictionary that contains the friendly/humanly readable titles for the nodes
@@ -49,5 +51,6 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "unloadAllModels": "卸载所有已加载模型",
     "TraverseFolder": "遍历文件夹",
     "LoadTextCustom": "读取文本文件(本地)",
-    "LoadTextCustomOnline": "读取文本文件(线上)"
+    "LoadTextCustomOnline": "读取文本文件(线上)",
+    "HeyGemF2F": "HeyGem口型同步(API)",
 }
diff --git a/nodes/heygem.py b/nodes/heygem.py
new file mode 100644
index 0000000..00fd998
--- /dev/null
+++ b/nodes/heygem.py
@@ -0,0 +1,92 @@
+import os
+import uuid
+
+import loguru
+import requests
+import torchaudio
+import torchvision
+from time import sleep
+from torch import Tensor
+
+
+class HeyGemF2F:
+    """HeyGem 嘴型同步"""
+
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "video": ("IMAGE", {"forceInput": True}),
+                "audio": ("AUDIO", {"forceInput": True}),
+                "heygem_url": ("STRING", {"default": "http://127.0.0.1:8383"}),
+                "heygem_temp_path": ("STRING", {"default":"/code/data/temp"})
+            }
+        }
+
+    RETURN_TYPES = ("STRING",)
+    RETURN_NAMES = ("视频存储路径",)
+    FUNCTION = "f2f"
+    CATEGORY = "不忘科技-自定义节点🚩"
+
+    def path_convert(self,path):
+        if ":" in path:
+            path = path.replace(os.sep,"/").split(":")
+            path[0] = path[0].lower()
+            path[1] = path[1][1:]
+            path = "/".join(["/mnt",*path])
+        return path
+
+    def f2f(self, video:Tensor, audio:dict, heygem_url:str, heygem_temp_path:str):
+        uid = str(uuid.uuid4())
+        try:
+            try:
+                torchvision.io.write_video(os.path.join(os.path.dirname(__file__),"%s.mp4" % uid), video.mul_(255).int(),25)
+                torchaudio.save(os.path.join(os.path.dirname(__file__),"%s.wav" % uid), audio["waveform"].squeeze(0), audio["sample_rate"], True)
+            except:
+                raise RuntimeError("Save Temp File Error! ")
+            payload = {
+                "code": uid,
+                "video_url": self.path_convert(os.path.join(os.path.dirname(__file__),"%s.mp4" % uid)),
+                "audio_url": self.path_convert(os.path.join(os.path.dirname(__file__),"%s.wav" % uid)),
+                "chaofen": 1,
+                "watermark_switch": 0,
+                "pn": 1
+            }
+            print(payload)
+            r = requests.post(heygem_url+"/easy/submit", json=payload)
+            if r.status_code != 200:
+                raise RuntimeError("Request Error!")
+            else:
+                r_json = r.json()
+                if r_json["success"]:
+                    loguru.logger.info("Submit Task Success")
+                else:
+                    raise RuntimeError("Submit Task Fail")
+            t = 30
+            while t>0:
+                r = requests.get(heygem_url+"/easy/query?code="+uid)
+                if r.status_code == 200:
+                    j = r.json()
+                    if "msg" in j and j["msg"]=="任务不存在":
+                        raise RuntimeError("Task Missing")
+                    if "data" in j and "异常" in j["data"]["msg"]:
+                        raise RuntimeError("Task Run Error: %s" % j["data"]["msg"])
+                    if "data" in j and j["data"]["progress"] < 100 and j["data"]["result"] == "":
+                        loguru.logger.info("Waiting Task Finish")
+                    elif "data" in j and j["data"]["progress"] == 100 and j["data"]["msg"] == "任务完成" and j["data"]["result"] != "":
+                        loguru.logger.info("Task Finished")
+                        return ("%s%s" % (heygem_temp_path, j["data"]["result"]),)
+                    else:
+                        loguru.logger.info("Unknown Status")
+                else:
+                    loguru.logger.info("Get Task Status Failed")
+                t -= 1
+                sleep(5)
+        except BaseException as e:
+            raise RuntimeError(e)
+        finally:
+            try:
+                os.remove(os.path.join(os.path.dirname(__file__),"%s.mp4" % uid))
+                os.remove(os.path.join(os.path.dirname(__file__),"%s.wav" % uid))
+            except:
+                pass
\ No newline at end of file