chore: sync code and project files

modules/script_gen.py

@@ -6,6 +6,7 @@ import base64
 import json
 import logging
 import os
+import time
 import requests
 from typing import Dict, Any, List, Optional
 from pathlib import Path
@@ -27,7 +28,10 @@ class ScriptGenerator:
         # OpenAI-compatible client for ShuBiaoBiao (supports multiple models incl. GPT)
         self.shubiaobiao_client = OpenAI(
             api_key=config.SHUBIAOBIAO_KEY,
-            base_url=config.SHUBIAOBIAO_BASE_URL
+            base_url=config.SHUBIAOBIAO_BASE_URL,
+            # IMPORTANT: OpenAI SDK default timeout is 10 minutes; cap it to keep UX responsive.
+            timeout=config.SHUBIAOBIAO_CHAT_TIMEOUT_S,
+            max_retries=config.SHUBIAOBIAO_CHAT_MAX_RETRIES,
         )
         
         # Default System Prompt
@@ -139,15 +143,23 @@ class ScriptGenerator:
         product_name: str, 
         product_info: Dict[str, Any], 
         image_paths: List[str] = None,
-        model_provider: str = "shubiaobiao" # "shubiaobiao" or "doubao"
+        model_provider: str = "shubiaobiao", # "shubiaobiao" or "doubao"
+        user_id: Optional[str] = None,
     ) -> Dict[str, Any]:
         """
         生成分镜脚本
         """
         logger.info(f"Generating script for: {product_name} (Provider: {model_provider})")
         
-        # 1. 构造 Prompt (优先从数据库读取配置)
-        system_prompt = db.get_config("prompt_script_gen", self.default_system_prompt)
+        # 1. 构造 Prompt (优先按 user_id 读取；否则回退到全局配置，再回退默认)
+        system_prompt = None
+        if user_id:
+            try:
+                system_prompt = db.get_user_prompt(user_id, "prompt_script_gen")
+            except Exception:
+                system_prompt = None
+        if not system_prompt:
+            system_prompt = db.get_config("prompt_script_gen", self.default_system_prompt)
         user_prompt = self._build_user_prompt(product_name, product_info)
         
         # Branch for Doubao (Volcengine)
@@ -293,21 +305,40 @@ class ScriptGenerator:
         ShuBiaoBiao OpenAI-compatible multimodal chat.
         IMPORTANT: For ShuBiaoBiao models, we pass image URLs (R2 public URLs), not base64.
         """
+        t0 = time.time()
+        # Use WARNING level so it shows up even if Streamlit/root logger is not configured to INFO.
+        logger.warning(
+            f"[script_gen] start shubiaobiao chat model={model_name} images={len(image_paths or [])} "
+            f"timeout_s={getattr(config, 'SHUBIAOBIAO_CHAT_TIMEOUT_S', 'n/a')} "
+            f"max_retries={getattr(config, 'SHUBIAOBIAO_CHAT_MAX_RETRIES', 'n/a')}"
+        )
         messages = [{"role": "system", "content": system_prompt}]
         user_content: List[Dict[str, Any]] = []
         # Images first (URL), then text
+        t_upload0 = time.time()
         urls = self._upload_images_to_r2(image_paths or [], limit=10)
+        logger.warning(
+            f"[script_gen] r2_upload done urls={len(urls)} elapsed_s={time.time() - t_upload0:.2f}"
+        )
         for url in urls:
             user_content.append({"type": "image_url", "image_url": {"url": url}})
         user_content.append({"type": "text", "text": user_prompt})
         messages.append({"role": "user", "content": user_content})
 
         try:
-            resp = self.shubiaobiao_client.chat.completions.create(
+            client = self.shubiaobiao_client.with_options(
+                timeout=config.SHUBIAOBIAO_CHAT_TIMEOUT_S,
+                max_retries=config.SHUBIAOBIAO_CHAT_MAX_RETRIES,
+            )
+            t_call0 = time.time()
+            resp = client.chat.completions.create(
                 model=model_name,
                 messages=messages,
                 temperature=0.7,
             )
+            logger.warning(
+                f"[script_gen] shubiaobiao chat done elapsed_s={time.time() - t_call0:.2f} total_s={time.time() - t0:.2f}"
+            )
             content_text = (resp.choices[0].message.content or "").strip()
             script_json = self._extract_json_from_response(content_text)
             if script_json is None:
@@ -323,7 +354,9 @@ class ScriptGenerator:
             }
             return final_script
         except Exception as e:
-            logger.error(f"shubiaobiao script generation failed ({model_name}): {e}")
+            logger.error(
+                f"shubiaobiao script generation failed ({model_name}) after {time.time() - t0:.2f}s: {e}"
+            )
             return None
 
     def _postprocess_selling_points(self, product_info: Dict[str, Any], selling_points: Any) -> List[str]:
@@ -582,7 +615,33 @@ class ScriptGenerator:
 
     def _validate_and_fix_script(self, script: Dict[str, Any]) -> Dict[str, Any]:
         """校验并修复脚本结构"""
-        # 简单校验，确保必要字段存在
-        if "scenes" not in script:
+        if not isinstance(script, dict):
+            return {"scenes": []}
+
+        # Ensure fields exist
+        if "scenes" not in script or not isinstance(script.get("scenes"), list):
             script["scenes"] = []
+
+        # Normalize: keep visual_anchor at top-level, but avoid repeating the full anchor in every scene.visual_prompt.
+        # Reason: repeating a long anchor 4-5 times explodes tokens and makes UI look like "only three sections",
+        # while image generation already supports passing visual_anchor separately and prepending it at runtime.
+        visual_anchor = script.get("visual_anchor") or ""
+        if isinstance(visual_anchor, str) and visual_anchor.strip() and script["scenes"]:
+            anchor = visual_anchor.strip()
+            prefix = f"[{anchor}]"
+            for scene in script["scenes"]:
+                if not isinstance(scene, dict):
+                    continue
+                vp = scene.get("visual_prompt")
+                if not isinstance(vp, str) or not vp.strip():
+                    continue
+                s = vp.strip()
+                # Strip exact "[anchor]" prefix if present
+                if s.startswith(prefix):
+                    s = s[len(prefix):].lstrip()
+                # If the model output copied the raw anchor without brackets, strip it too
+                elif s.startswith(anchor):
+                    s = s[len(anchor):].lstrip()
+                scene["visual_prompt"] = s
+
         return script