fix: 文生图、图生图支持 GPT

.vscode/launch.json

@@ -5,20 +5,9 @@
       "name": "text-to-image",
       "type": "debugpy",
       "request": "launch",
-      "program": "skills/text-to-image/scripts/text_to_image.py",
+      "program": "skills/text-to-image/scripts/debug_openai_image_generation_test.py",
       "console": "integratedTerminal",
       "justMyCode": true,
-      "args": [
-        "{\"prompt\":\"一只站在雨夜街头的白猫\",\"model\":\"jimeng-5.0\",\"negative_prompt\":\"模糊, 低清\",\"ratio\":\"16:9\",\"resolution\":\"2k\"}"
-      ],
-      "env": {
-        "ROBOT_FROM_WX_ID": "57004904192@chatroom",
-        "ROBOT_CODE": "houhouipad",
-        "MYSQL_HOST": "127.0.0.1",
-        "MYSQL_PORT": "3306",
-        "MYSQL_USER": "root",
-        "MYSQL_PASSWORD": "houhou"
-      }
     }
   ]
 }

skills/image-to-image/scripts/image_to_image.py

@@ -236,50 +236,57 @@ def _openai_client(config: dict) -> OpenAI:
     if not api_key:
         raise RuntimeError("OpenAI 绘图配置缺少 api_key")
 
-    kwargs: dict[str, str | float] = {"api_key": api_key}
     base_url = str(config.get("base_url", "") or "").strip()
-    if base_url:
-        kwargs["base_url"] = base_url
     organization = str(config.get("organization", "") or "").strip()
-    if organization:
-        kwargs["organization"] = organization
     project = str(config.get("project", "") or "").strip()
-    if project:
+    timeout: float | None = None
-        kwargs["project"] = project
     timeout_value = config.get("timeout")
     if timeout_value not in (None, ""):
-        kwargs["timeout"] = float(timeout_value)
+        timeout = float(timeout_value)
 
-    return OpenAI(**kwargs)
+    return OpenAI(
+        api_key=api_key,
+        base_url=base_url or None,
+        organization=organization or None,
+        project=project or None,
+        timeout=timeout,
+    )
 
 
-def _openai_image_suffix(output_format: str) -> str:
+def _truncate_debug_payload(value):
-    if output_format == "jpeg":
+    if isinstance(value, dict):
-        return ".jpg"
+        return {
-    return f".{output_format}"
+            key: (
+                f"{item[:50]}..." if key == "b64_json" and isinstance(item, str) and len(item) > 50 else _truncate_debug_payload(item)
+            )
+            for key, item in value.items()
+        }
+    if isinstance(value, list):
+        return [_truncate_debug_payload(item) for item in value]
+    return value
 
 
-def _write_openai_image_file(b64_json: str, output_format: str) -> str:
+def _debug_response(label: str, payload) -> None:
-    image_bytes = base64.b64decode(b64_json)
+    if hasattr(payload, "model_dump"):
-    with tempfile.NamedTemporaryFile(
+        payload = payload.model_dump()
-        prefix="wechat-openai-image-",
+    payload = _truncate_debug_payload(payload)
-        suffix=_openai_image_suffix(output_format),
+    sys.stdout.write(f"[debug] {label}: {json.dumps(payload, ensure_ascii=False)}\n")
-        delete=False,
-    ) as image_file:
-        image_file.write(image_bytes)
-        return image_file.name
 
 
-def _openai_images_from_response(response, output_format: str) -> list[str]:
+def _rewrite_openai_image_url(url: str) -> str:
+    internal_host = "http://chatgpt2api:80"
+    external_host = "http://chatgpt2api.houhoukang.com"
+    if url.startswith(internal_host):
+        return f"{external_host}{url[len(internal_host):]}"
+    return url
+
+
+def _openai_images_from_response(response) -> list[str]:
     outputs: list[str] = []
     for item in getattr(response, "data", []) or []:
-        b64_json = getattr(item, "b64_json", None)
-        if b64_json:
-            outputs.append(_write_openai_image_file(b64_json, output_format))
-            continue
         url = getattr(item, "url", None)
         if url:
-            outputs.append(url)
+            outputs.append(_rewrite_openai_image_url(str(url)))
     return outputs
 
 
@@ -297,7 +304,8 @@ def _send_image_outputs(client_port: str, from_wx_id: str, image_outputs: list[s
             "to_wxid": from_wx_id,
             "image_urls": remote_urls,
         }
-        _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        response = _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        _debug_response("send image url response", response)
 
     if local_paths:
         send_url = f"http://127.0.0.1:{client_port}/api/v1/robot/message/send/image/local"
@@ -305,7 +313,8 @@ def _send_image_outputs(client_port: str, from_wx_id: str, image_outputs: list[s
             "to_wxid": from_wx_id,
             "file_path": local_paths,
         }
-        _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        response = _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        _debug_response("send image local response", response)
 
 
 def _cleanup_openai_temp_files(image_outputs: list[str]) -> None:
@@ -531,7 +540,8 @@ def call_openai(config: dict, prompt: str, model: str, images: list[str],
             for input_file in input_files:
                 input_file.close()
 
-    return _openai_images_from_response(response, output_format)
+    _debug_response("openai images.edit response", response)
+    return _openai_images_from_response(response)
 
 
 # ---------------------------------------------------------------------------

skills/text-to-image/scripts/text_to_image.py

@@ -3,13 +3,11 @@
 from __future__ import annotations
 
 import argparse
-import base64
 import json
 import os
 import re
 import subprocess
 import sys
-import tempfile
 import time
 import traceback
 import urllib.parse
@@ -237,50 +235,57 @@ def _openai_client(config: dict) -> OpenAI:
     if not api_key:
         raise RuntimeError("OpenAI 绘图配置缺少 api_key")
 
-    kwargs: dict[str, str | float] = {"api_key": api_key}
     base_url = str(config.get("base_url", "") or "").strip()
-    if base_url:
-        kwargs["base_url"] = base_url
     organization = str(config.get("organization", "") or "").strip()
-    if organization:
-        kwargs["organization"] = organization
     project = str(config.get("project", "") or "").strip()
-    if project:
+    timeout: float | None = None
-        kwargs["project"] = project
     timeout_value = config.get("timeout")
     if timeout_value not in (None, ""):
-        kwargs["timeout"] = float(timeout_value)
+        timeout = float(timeout_value)
 
-    return OpenAI(**kwargs)
+    return OpenAI(
+        api_key=api_key,
+        base_url=base_url or None,
+        organization=organization or None,
+        project=project or None,
+        timeout=timeout,
+    )
 
 
-def _openai_image_suffix(output_format: str) -> str:
+def _truncate_debug_payload(value):
-    if output_format == "jpeg":
+    if isinstance(value, dict):
-        return ".jpg"
+        return {
-    return f".{output_format}"
+            key: (
+                f"{item[:50]}..." if key == "b64_json" and isinstance(item, str) and len(item) > 50 else _truncate_debug_payload(item)
+            )
+            for key, item in value.items()
+        }
+    if isinstance(value, list):
+        return [_truncate_debug_payload(item) for item in value]
+    return value
 
 
-def _write_openai_image_file(b64_json: str, output_format: str) -> str:
+def _debug_response(label: str, payload) -> None:
-    image_bytes = base64.b64decode(b64_json)
+    if hasattr(payload, "model_dump"):
-    with tempfile.NamedTemporaryFile(
+        payload = payload.model_dump()
-        prefix="wechat-openai-image-",
+    payload = _truncate_debug_payload(payload)
-        suffix=_openai_image_suffix(output_format),
+    sys.stdout.write(f"[debug] {label}: {json.dumps(payload, ensure_ascii=False)}\n")
-        delete=False,
-    ) as image_file:
-        image_file.write(image_bytes)
-        return image_file.name
 
 
-def _openai_images_from_response(response, output_format: str) -> list[str]:
+def _rewrite_openai_image_url(url: str) -> str:
+    internal_host = "http://chatgpt2api:80"
+    external_host = "http://chatgpt2api.houhoukang.com"
+    if url.startswith(internal_host):
+        return f"{external_host}{url[len(internal_host):]}"
+    return url
+
+
+def _openai_images_from_response(response) -> list[str]:
     outputs: list[str] = []
     for item in getattr(response, "data", []) or []:
-        b64_json = getattr(item, "b64_json", None)
-        if b64_json:
-            outputs.append(_write_openai_image_file(b64_json, output_format))
-            continue
         url = getattr(item, "url", None)
         if url:
-            outputs.append(url)
+            outputs.append(_rewrite_openai_image_url(str(url)))
     return outputs
 
 
@@ -298,7 +303,8 @@ def _send_image_outputs(client_port: str, from_wx_id: str, image_outputs: list[s
             "to_wxid": from_wx_id,
             "image_urls": remote_urls,
         }
-        _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        response = _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        _debug_response("send image url response", response)
 
     if local_paths:
         send_url = f"http://127.0.0.1:{client_port}/api/v1/robot/message/send/image/local"
@@ -306,7 +312,8 @@ def _send_image_outputs(client_port: str, from_wx_id: str, image_outputs: list[s
             "to_wxid": from_wx_id,
             "file_path": local_paths,
         }
-        _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        response = _http_post_json(send_url, send_body, {"Content-Type": "application/json"}, timeout=60)
+        _debug_response("send image local response", response)
 
 
 def _cleanup_openai_temp_files(image_outputs: list[str]) -> None:
@@ -490,7 +497,8 @@ def call_openai(config: dict, prompt: str, model: str,
         kwargs["output_compression"] = _coerce_int(config.get("output_compression"), 100, 0, 100)
 
     response = client.images.generate(**kwargs)
-    return _openai_images_from_response(response, output_format)
+    _debug_response("openai images.generate response", response)
+    return _openai_images_from_response(response)
 
 
 # ---------------------------------------------------------------------------