增加了图片目录模糊匹配

poster_gen_config.json

@@ -1,7 +1,7 @@
 {
     "date": "4月29日,4月30日, 4月28日, 5月1日",
-    "num": 10,
+    "num": 5,
-    "variants": 5,
+    "variants": 1,
     "topic_temperature": 0.2,
     "topic_top_p": 0.3,
     "topic_presence_penalty": 1.5,
@@ -42,13 +42,13 @@
         {
             "type": "Object",
             "file_path": [
-                "./resource/Object/笔架山居森林度假酒店.txt"
+                "./resource/Object/美的鹭湖鹭栖台酒店.txt"
             ]
         },
         {
             "type": "Description",
             "file_path": [
-                "./resource/Object/笔架山居森林度假酒店.txt"
+                "./resource/Object/美的鹭湖鹭栖台酒店.txt"
             ]
         },
         {

utils/tweet_generator.py

@@ -10,6 +10,7 @@ from datetime import datetime
 import sys
 import traceback
 import logging # Add logging
+import re
 # sys.path.append('/root/autodl-tmp') # No longer needed if running as a module or if path is set correctly
 # 从本地模块导入
 # from TravelContentCreator.core.ai_agent import AI_Agent # Remove project name prefix
@@ -548,13 +549,88 @@ def generate_posters_for_topic(topic_item: dict,
     # Construct and check INPUT image paths
     input_img_dir_path = os.path.join(image_base_dir, object_name)
     if not os.path.exists(input_img_dir_path) or not os.path.isdir(input_img_dir_path):
-        logging.warning(f"Warning: Modify Image directory not found or not a directory: '{input_img_dir_path}'. Skipping posters for this topic.")
+        # 模糊匹配：如果找不到完全匹配的目录，尝试查找包含关键词的目录
-        return False
+        logging.info(f"尝试对图片目录进行模糊匹配: {object_name}")
+        found_dir = None
+        
+        # 1. 尝试获取image_base_dir下的所有目录
+        try:
+            all_dirs = [d for d in os.listdir(image_base_dir) 
+                       if os.path.isdir(os.path.join(image_base_dir, d))]
+            logging.info(f"找到 {len(all_dirs)} 个图片目录可用于模糊匹配")
+            
+            # 2. 提取对象名称中的关键词
+            # 例如："美的鹭湖鹭栖台酒店+盈香心动乐园" -> ["美的", "鹭湖", "酒店", "乐园"]
+            # 首先通过常见分隔符分割（+、空格、_、-等）
+            parts = re.split(r'[+\s_\-]', object_name)
+            keywords = []
+            for part in parts:
+                # 只保留长度大于1的有意义关键词
+                if len(part) > 1:
+                    keywords.append(part)
+            
+            # 尝试匹配更短的语义单元（例如中文的2-3个字的词语）
+            # 对于中文名称，可以尝试提取2-3个字的短语
+            for i in range(len(object_name) - 1):
+                keyword = object_name[i:i+2]  # 提取2个字符
+                if len(keyword) == 2 and all('\u4e00' <= c <= '\u9fff' for c in keyword):
+                    keywords.append(keyword)
+            
+            # 3. 对每个目录进行评分
+            dir_scores = {}
+            for directory in all_dirs:
+                score = 0
+                dir_lower = directory.lower()
+                # 为每个匹配的关键词增加分数
+                for keyword in keywords:
+                    if keyword.lower() in dir_lower:
+                        score += 1
+                
+                # 如果得分大于0（至少匹配一个关键词），记录该目录
+                if score > 0:
+                    dir_scores[directory] = score
+            
+            # 4. 选择得分最高的目录
+            if dir_scores:
+                best_match = max(dir_scores.items(), key=lambda x: x[1])
+                found_dir = best_match[0]
+                logging.info(f"模糊匹配成功！匹配目录: {found_dir}，匹配分数: {best_match[1]}")
+                
+                # 更新图片目录路径
+                input_img_dir_path = os.path.join(image_base_dir, found_dir)
+                logging.info(f"使用模糊匹配的图片目录: {input_img_dir_path}")
+            else:
+                logging.warning(f"模糊匹配未找到任何包含关键词的目录")
+        except Exception as e:
+            logging.warning(f"模糊匹配过程中出错: {e}")
+            
+        # 如果仍然无法找到有效目录，则返回错误
+        if not found_dir or not os.path.exists(input_img_dir_path) or not os.path.isdir(input_img_dir_path):
+            logging.warning(f"Warning: 即使通过模糊匹配也无法找到图片目录: '{input_img_dir_path}'. Skipping posters for this topic.")
+            return False
 
     # Locate Description File using resource_dir_config parameter
     info_directory = []
     description_file_path = None
     found_description = False
+    
+    # 准备关键词列表用于模糊匹配
+    # 与上面图片目录匹配类似，提取对象名称的关键词
+    parts = re.split(r'[+\s_\-]', object_name)
+    keywords = []
+    for part in parts:
+        if len(part) > 1:
+            keywords.append(part)
+    
+    # 尝试提取中文短语作为关键词
+    for i in range(len(object_name) - 1):
+        keyword = object_name[i:i+2]
+        if len(keyword) == 2 and all('\u4e00' <= c <= '\u9fff' for c in keyword):
+            keywords.append(keyword)
+    
+    logging.info(f"用于描述文件模糊匹配的关键词: {keywords}")
+    
+    # 尝试精确匹配
     for dir_info in resource_dir_config:
         if dir_info.get("type") == "Description":
             for file_path in dir_info.get("file_path", []):
@@ -562,15 +638,45 @@ def generate_posters_for_topic(topic_item: dict,
                     description_file_path = file_path
                     if os.path.exists(description_file_path):
                         info_directory = [description_file_path]
-                        logging.info(f"Found and using description file from config: {description_file_path}")
+                        logging.info(f"找到并使用精确匹配的描述文件: {description_file_path}")
                         found_description = True
                     else:
-                        logging.warning(f"Warning: Description file specified in config not found: {description_file_path}")
+                        logging.warning(f"Warning: 配置中指定的描述文件未找到: {description_file_path}")
                     break
             if found_description:
-                break 
+                break
+    
+    # 如果精确匹配失败，尝试模糊匹配
     if not found_description:
-        logging.info(f"Warning: No matching description file found for object '{object_name}' in config resource_dir (type='Description').")
+        logging.info(f"未找到'{object_name}'的精确匹配描述文件，尝试模糊匹配...")
+        best_score = 0
+        best_file = None
+        
+        for dir_info in resource_dir_config:
+            if dir_info.get("type") == "Description":
+                for file_path in dir_info.get("file_path", []):
+                    file_name = os.path.basename(file_path)
+                    score = 0
+                    
+                    # 计算关键词匹配分数
+                    for keyword in keywords:
+                        if keyword.lower() in file_name.lower():
+                            score += 1
+                    
+                    # 如果当前文件得分更高，更新最佳匹配
+                    if score > best_score and os.path.exists(file_path):
+                        best_score = score
+                        best_file = file_path
+        
+        # 如果找到了最佳匹配文件
+        if best_file:
+            description_file_path = best_file
+            info_directory = [description_file_path]
+            logging.info(f"模糊匹配找到描述文件: {description_file_path}，匹配分数: {best_score}")
+            found_description = True
+    
+    if not found_description:
+        logging.warning(f"未找到对象'{object_name}'的匹配描述文件。")
     
     # Generate Text Configurations for All Variants
     try: