Fix: Use user preset inputs instead of random generation

Core changes:
- User inputs are now directly used (not regenerated)
- Missing fields are generated based on provided inputs
- Added convert_web_profile_to_selector_format() with clear logging
- Fixed data type issues in web_api_bridge.py
- Preserved user_profile.json to avoid overwriting preset values

The system now follows 'use what user provides, generate what's missing' principle.
Console output shows ✓ for user inputs and ✗ for generated fields.

BUGFIX_SUMMARY.md

@@ -0,0 +1,169 @@
+# Bug Fix Summary - Profile Generation Not Using Preset Inputs
+
+## 问题描述
+当用户通过Web界面输入部分预设内容（如年龄、性别、职业等）时，生成的profile仍然是完全随机的，没有使用用户的预设输入。
+
+## 用户需求
+**核心原则：用户输入了什么就用什么，没输入的才基于已输入内容生成**
+- 用户输入的字段：直接使用，不调用GPT重新生成
+- 用户未输入的字段：基于已输入的字段，调用 `based_data.py` 中的函数生成
+- 有多少用多少，灵活处理部分输入的情况
+
+## 根本原因
+在 `generate_profile.py` 的 `generate_single_profile()` 函数中，代码会调用 `select_attributes.py` 中的 `generate_user_profile()` 函数来重新生成用户配置文件。这个函数会**完全随机生成**所有基础信息（年龄、性别、职业、地点等），从而覆盖了用户在 `web_api_bridge.py` 中预设的输入。
+
+### 问题代码流程：
+1. 用户在Web界面输入预设值
+2. `web_api_bridge.py` 将预设值保存到 `user_profile.json`
+3. `generate_single_profile()` 被调用
+4. **问题点**: `generate_single_profile()` 调用 `generate_user_profile()` 重新随机生成配置
+5. 新生成的随机配置覆盖了用户的预设输入
+
+## 修复方案
+
+### 1. 修改 `generate_profile.py` (主要修改)
+**文件**: `/Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/code/generate_profile.py`
+
+**修改内容**:
+- 在调用属性选择之前，检查是否已存在 `user_profile.json`（由 `web_api_bridge.py` 保存）
+- 如果存在，读取并使用该文件（包含用户预设输入）
+- 将 web 格式的配置转换为 `AttributeSelector` 期望的格式
+- 在属性选择完成后，恢复原始的 `user_profile.json`，避免被覆盖
+
+**关键代码**:
+```python
+# 检查是否已经存在user_profile.json（由web_api_bridge.py保存）
+if os.path.exists(user_profile_path):
+    # 读取已存在的用户配置文件（包含预设输入）
+    print(f"Found existing user_profile.json, using preset inputs...")
+    with open(user_profile_path, 'r', encoding='utf-8') as f:
+        base_profile = json.load(f)
+    
+    # 保存原始配置，稍后恢复
+    original_base_profile = base_profile.copy()
+    using_preset = True
+    
+    # 转换为AttributeSelector期望的格式
+    user_profile = convert_web_profile_to_selector_format(base_profile)
+else:
+    # 如果不存在，生成新的配置文件
+    user_profile = gen_profile()
+```
+
+### 2. 添加格式转换函数到 `select_attributes.py`
+**文件**: `/Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/code/select_attributes.py`
+
+**新增函数**: `convert_web_profile_to_selector_format()`
+
+**功能**: 将 `web_api_bridge.py` 保存的配置格式转换为 `AttributeSelector` 期望的格式
+
+**核心逻辑**（严格遵循"有多少用多少"原则）:
+```python
+# 对每个字段进行检查：
+# 1. 如果用户提供了该字段 → 直接使用，打印 "✓ Using user-provided ..."
+# 2. 如果用户没提供 → 基于已有信息生成，打印 "✗ ... not provided, generated ..."
+
+# 示例：年龄处理
+age = web_profile.get("age")
+if age:
+    age_info = {"age": age, "age_group": calculate_age_group(age)}
+    print(f"✓ Using user-provided age: {age}")
+else:
+    age_info = generate_age_info()  # 只有没提供时才生成
+    print(f"✗ Age not provided, generated: {age_info['age']}")
+```
+
+**格式对比**:
+- Web格式: `{"age": int, "gender": str, "Occupations": [str], ...}`
+- Selector格式: `{"age_info": {"age": int, "age_group": str}, "gender": str, "career_info": {"status": str}, ...}`
+
+**处理的字段**:
+- ✅ age → age_info (如果提供了就用，否则生成)
+- ✅ gender (如果提供了就用，否则生成)
+- ✅ location (如果提供了就用，否则生成)
+- ✅ Occupations → career_info (如果提供了就用，否则生成)
+- ✅ personal_values (如果提供了就用，否则基于已有信息生成)
+- ✅ life_attitude (如果提供了就用，否则基于已有信息生成)
+- ✅ personal_story (如果提供了就用，否则基于已有信息生成)
+- ✅ interests (如果提供了就用，否则基于故事生成)
+
+### 3. 修复 `web_api_bridge.py` 中的数据类型问题
+**文件**: `/Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/code/web_api_bridge.py`
+
+**修复内容**:
+1. **修复 `values_orientation` 类型问题**:
+   - `generate_personal_values()` 返回字典，需要提取 `values_orientation` 字段
+   
+2. **修复 `personal_story` 嵌套问题**:
+   - 避免重复嵌套 `{"personal_story": {"personal_story": ...}}`
+
+## 修改文件列表
+1. ✅ `/Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/code/generate_profile.py`
+2. ✅ `/Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/code/select_attributes.py`
+3. ✅ `/Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/code/web_api_bridge.py`
+
+## 测试建议
+
+### 测试场景 1：部分输入（只输入基本信息）
+**输入**:
+- Age: 25
+- Gender: Male
+- Occupation: Software Engineer
+- City: Tokyo
+- Country: Japan
+
+**预期输出**:
+```
+✓ Using user-provided age: 25
+✓ Using user-provided gender: Male
+✓ Using user-provided location: Tokyo, Japan
+✓ Using user-provided occupation: Software Engineer
+✗ Personal values not provided, generated based on user inputs
+✗ Life attitude not provided, generated based on user inputs
+✗ Life story not provided, generated based on user inputs
+✗ Interests not provided, generated based on life story
+```
+
+### 测试场景 2：完整输入（输入所有字段）
+**输入**:
+- Age: 30
+- Gender: Female
+- Occupation: Teacher
+- Location: Paris, France
+- Personal Values: "Values education and creativity"
+- Life Attitude: "Optimistic and passionate about learning"
+- Life Story: "Born in Lyon, moved to Paris for university..."
+- Interests: "Reading, painting, traveling"
+
+**预期输出**:
+```
+✓ Using user-provided age: 30
+✓ Using user-provided gender: Female
+✓ Using user-provided location: Paris, France
+✓ Using user-provided occupation: Teacher
+✓ Using user-provided personal values: Values education and creativity
+✓ Using user-provided life attitude: Optimistic and passionate about learning
+✓ Using user-provided life story: Born in Lyon, moved to Paris for university...
+✓ Using user-provided interests: Reading, painting, traveling
+```
+
+### 测试场景 3：空输入（完全随机生成）
+**输入**: 全部留空
+
+**预期输出**:
+```
+✗ Age not provided, generated: 42
+✗ Gender not provided, generated: female
+✗ Location not provided, generated: Mumbai, India
+✗ Occupation not provided, generated: Marketing Manager
+✗ Personal values not provided, generated based on user inputs
+✗ Life attitude not provided, generated based on user inputs
+✗ Life story not provided, generated based on user inputs
+✗ Interests not provided, generated based on life story
+```
+
+## 预期结果
+- ✅ 用户输入的预设值会被**直接使用**，不会被重新生成
+- ✅ 未输入的字段会基于已输入的信息**智能生成**
+- ✅ 生成的profile与用户输入保持**完全一致性**
+- ✅ 控制台会清晰显示哪些字段使用了用户输入（✓），哪些是生成的（✗）

TEST_EXAMPLE.md

@@ -0,0 +1,121 @@
+# 测试示例 - 验证用户输入是否被正确使用
+
+## 如何测试
+
+### 方法 1: 通过 Web 界面测试
+
+1. 启动应用：
+```bash
+cd /Users/yufan_zhou/Documents/deeppersona/deeppersona-experience
+python app.py
+```
+
+2. 在浏览器中打开 `http://localhost:7860`
+
+3. 输入测试数据（例如）：
+   - **Age**: 28
+   - **Gender**: Male
+   - **Occupation**: Data Scientist
+   - **City**: San Francisco
+   - **Country**: USA
+   - **Personal Values**: (留空，测试自动生成)
+   - **Life Attitude**: "Curious and analytical"
+   - **Life Story**: (留空，测试自动生成)
+   - **Interests and Hobbies**: "Machine learning, hiking"
+
+4. 点击 "Generate Character Profile"
+
+5. **检查控制台输出**，应该看到类似：
+```
+Found existing user_profile.json, using preset inputs...
+✓ Using user-provided age: 28
+✓ Using user-provided gender: Male
+✓ Using user-provided location: San Francisco, USA
+✓ Using user-provided occupation: Data Scientist
+✗ Personal values not provided, generated based on user inputs
+✓ Using user-provided life attitude: Curious and analytical
+✗ Life story not provided, generated based on user inputs
+✓ Using user-provided interests: Machine learning, hiking
+```
+
+6. **检查生成的 Profile**，应该包含：
+   - 年龄 28 岁
+   - 性别 Male
+   - 职业 Data Scientist
+   - 地点 San Francisco, USA
+   - 生活态度包含 "Curious and analytical"
+   - 兴趣包含 "Machine learning" 和 "hiking"
+
+### 方法 2: 通过命令行测试
+
+1. 创建测试输入文件 `test_input.json`:
+```json
+{
+  "basic_info": {
+    "age": 25,
+    "gender": "Female",
+    "occupation": {
+      "status": "Software Engineer"
+    },
+    "location": {
+      "city": "Tokyo",
+      "country": "Japan"
+    }
+  },
+  "custom_values": {
+    "personal_values": "Values innovation and teamwork",
+    "life_attitude": "",
+    "life_story": "",
+    "interests_hobbies": "Coding, anime, gaming"
+  }
+}
+```
+
+2. 运行测试：
+```bash
+cd /Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/code
+python web_api_bridge.py --input test_input.json --attributes 200
+```
+
+3. **检查输出**，应该显示：
+```
+✓ Using user-provided age: 25
+✓ Using user-provided gender: Female
+✓ Using user-provided location: Tokyo, Japan
+✓ Using user-provided occupation: Software Engineer
+✓ Using user-provided personal values: Values innovation and teamwork
+✗ Life attitude not provided, generated based on user inputs
+✗ Life story not provided, generated based on user inputs
+✓ Using user-provided interests: Coding, anime, gaming
+```
+
+## 验证要点
+
+### ✅ 正确行为
+1. 用户输入的字段显示 `✓ Using user-provided ...`
+2. 未输入的字段显示 `✗ ... not provided, generated ...`
+3. 生成的 profile 包含所有用户输入的内容
+4. 未输入的字段基于已输入内容合理生成（不是完全随机）
+
+### ❌ 错误行为（如果看到这些，说明修复失败）
+1. 所有字段都显示 `✗ ... not provided, generated ...`（说明用户输入被忽略）
+2. 用户输入的年龄、性别等与生成的 profile 不匹配
+3. 生成的内容与用户输入完全无关
+
+## 调试提示
+
+如果测试失败，检查以下内容：
+
+1. **检查 `user_profile.json` 是否正确保存**:
+```bash
+cat /Users/yufan_zhou/Documents/deeppersona/deeppersona-experience/generate_user_profile_final/output/user_profile.json
+```
+
+2. **检查控制台是否显示 "Found existing user_profile.json"**:
+   - 如果没有，说明文件路径不对
+   - 如果显示 "No existing user_profile.json found"，说明文件没有被保存
+
+3. **检查是否有错误信息**:
+   - 查看完整的 traceback
+   - 检查是否有 JSON 解析错误
+   - 检查是否有字段类型不匹配的错误

generate_user_profile_final/code/generate_profile.py

@@ -372,28 +372,71 @@ def generate_single_profile(template: Dict = None, profile_index: int = 0, attri
     """
 
     
-    # First, run select_attributes.py to update base files (user_profile.json and selected_paths.json)
-    print(f'Running select_attributes.py to update base files with {attribute_count} attributes...')
+    # First, check if user_profile.json already exists (from web_api_bridge.py)
+    # If it exists, use it; otherwise generate a new one
+    print(f'Preparing to select {attribute_count} attributes...')
     try:
         # 直接导入select_attributes模块的函数，而不是通过subprocess运行
         import sys
         import os
         sys.path.append(os.path.dirname(os.path.abspath(__file__)))
-        from select_attributes import generate_user_profile as gen_profile
-        from select_attributes import get_selected_attributes, save_results
+        from select_attributes import get_selected_attributes, save_results, convert_web_profile_to_selector_format
+        
+        # 检查是否已经存在user_profile.json（由web_api_bridge.py保存）
+        correct_output_dir = os.path.join(get_project_root(), "output")
+        user_profile_path = os.path.join(correct_output_dir, 'user_profile.json')
+        
+        # 标记是否使用了预设配置
+        using_preset = False
+        original_base_profile = None
+        
+        if os.path.exists(user_profile_path):
+            # 读取已存在的用户配置文件（包含预设输入）
+            print(f"Found existing user_profile.json, using preset inputs...")
+            with open(user_profile_path, 'r', encoding='utf-8') as f:
+                base_profile = json.load(f)
+            
+            # 保存原始配置，稍后恢复
+            original_base_profile = base_profile.copy()
+            using_preset = True
+            
+            # 转换为AttributeSelector期望的格式
+            user_profile = convert_web_profile_to_selector_format(base_profile)
+            print(f"Using preset profile with age={user_profile.get('age_info', {}).get('age')}, gender={user_profile.get('gender')}")
+        else:
+            # 如果不存在，生成新的配置文件
+            print("No existing user_profile.json found, generating new random profile...")
+            from select_attributes import generate_user_profile as gen_profile
+            user_profile = gen_profile()
         
-        # 生成用户配置文件
-        user_profile = gen_profile()
         # 获取指定数量的属性
         selected_paths = get_selected_attributes(user_profile, attribute_count)
-        # 保存结果
-        correct_output_dir = os.path.join(get_project_root(), "output")
-        save_results(user_profile, selected_paths, correct_output_dir)
+        
+        # 只保存selected_paths，不覆盖user_profile.json
+        # 如果使用了预设配置，需要恢复原始文件
+        if using_preset and original_base_profile:
+            # 恢复原始的user_profile.json（web格式）
+            with open(user_profile_path, 'w', encoding='utf-8') as f:
+                json.dump(original_base_profile, f, ensure_ascii=False, indent=2)
+            print(f"Restored original user_profile.json with preset inputs")
+        else:
+            # 如果是新生成的，保存完整结果
+            save_results(user_profile, selected_paths, correct_output_dir)
+        
+        # 无论哪种情况，都要保存selected_paths
+        from select_attributes import build_nested_dict
+        nested_selected_paths = build_nested_dict(selected_paths)
+        paths_path = os.path.join(correct_output_dir, "selected_paths.json")
+        with open(paths_path, 'w', encoding='utf-8') as f:
+            json.dump(nested_selected_paths, f, ensure_ascii=False, indent=2)
+        print(f"Saved selected_paths.json with {len(selected_paths)} attributes")
         
         # 复制文件从源位置到目标位置
         copy_files_from_source_to_target()
     except Exception as e:
         print(f"Error executing select_attributes functions: {e}")
+        import traceback
+        traceback.print_exc()
         return {}
 
     # Load basic profile information and selected paths (base info is only a reference for GPT generation)

generate_user_profile_final/code/select_attributes.py

@@ -772,6 +772,199 @@ class AttributeSelector:
     
 
 
+def convert_web_profile_to_selector_format(web_profile: Dict) -> Dict:
+    """
+    将web_api_bridge.py保存的配置文件格式转换为AttributeSelector期望的格式
+    
+    核心原则：用户输入了什么就用什么，没输入的才生成
+    
+    web_profile格式:
+    {
+        "age": int,
+        "gender": str,
+        "Occupations": [str],
+        "location": {"city": str, "country": str},
+        "personal_values": {"values_orientation": str},
+        "life_attitude": str or dict,
+        "personal_story": {"personal_story": str},
+        "interests": {"interests": [str]}
+    }
+    
+    selector格式:
+    {
+        "age_info": {"age": int, "age_group": str},
+        "gender": str,
+        "location": {"city": str, "country": str},
+        "career_info": {"status": str},
+        "personal_values": {"values_orientation": str},
+        "life_attitude": {"attitude": str, "attitude_details": str, "coping_mechanism": str},
+        "personal_story": {"personal_story": str},
+        "interests": {"interests": [str]}
+    }
+    """
+    # 提取年龄信息 - 如果用户提供了就用，否则生成
+    age = web_profile.get("age")
+    if age:
+        # 根据年龄确定年龄组
+        if age <= 6:
+            age_group = "toddler"
+        elif age <= 12:
+            age_group = "child"
+        elif age <= 19:
+            age_group = "adolescent"
+        elif age <= 29:
+            age_group = "young_adult"
+        elif age <= 45:
+            age_group = "adult"
+        elif age <= 65:
+            age_group = "middle_aged"
+        else:
+            age_group = "senior"
+        
+        age_info = {"age": age, "age_group": age_group}
+        print(f"✓ Using user-provided age: {age}")
+    else:
+        age_info = generate_age_info()
+        print(f"✗ Age not provided, generated: {age_info['age']}")
+    
+    # 提取性别 - 如果用户提供了就用，否则生成
+    gender = web_profile.get("gender")
+    if gender:
+        print(f"✓ Using user-provided gender: {gender}")
+    else:
+        gender = generate_gender()
+        print(f"✗ Gender not provided, generated: {gender}")
+    
+    # 提取位置 - 如果用户提供了就用，否则生成
+    location = web_profile.get("location")
+    if location and location.get("city") and location.get("country"):
+        print(f"✓ Using user-provided location: {location.get('city')}, {location.get('country')}")
+    else:
+        location = generate_location()
+        print(f"✗ Location not provided, generated: {location.get('city')}, {location.get('country')}")
+    
+    # 提取职业信息 - 如果用户提供了就用，否则生成
+    occupations = web_profile.get("Occupations", [])
+    if occupations and len(occupations) > 0 and occupations[0]:
+        career_info = {"status": occupations[0]}
+        print(f"✓ Using user-provided occupation: {occupations[0]}")
+    else:
+        career_info = generate_career_info(age_info["age"])
+        print(f"✗ Occupation not provided, generated: {career_info['status']}")
+    
+    # 提取个人价值观 - 如果用户提供了就用，否则生成
+    personal_values = web_profile.get("personal_values", {})
+    values_orientation = personal_values.get("values_orientation", "")
+    if values_orientation and values_orientation.strip():
+        # 用户提供了价值观，直接使用
+        personal_values = {"values_orientation": values_orientation.strip()}
+        print(f"✓ Using user-provided personal values: {values_orientation[:50]}...")
+    else:
+        # 用户没提供，基于已有信息生成
+        personal_values = generate_personal_values(
+            age=age_info["age"],
+            gender=gender,
+            occupation=career_info["status"],
+            location=location
+        )
+        print(f"✗ Personal values not provided, generated based on user inputs")
+    
+    # 提取生活态度 - 如果用户提供了就用，否则生成
+    life_attitude_data = web_profile.get("life_attitude")
+    if life_attitude_data:
+        if isinstance(life_attitude_data, str) and life_attitude_data.strip():
+            # 用户提供了字符串格式的生活态度
+            life_attitude = {
+                "attitude": life_attitude_data.strip(),
+                "attitude_details": "",
+                "coping_mechanism": ""
+            }
+            print(f"✓ Using user-provided life attitude: {life_attitude_data[:50]}...")
+        elif isinstance(life_attitude_data, dict) and life_attitude_data.get("attitude"):
+            # 用户提供了字典格式的生活态度
+            life_attitude = life_attitude_data
+            print(f"✓ Using user-provided life attitude (dict format)")
+        else:
+            # 数据格式不对或为空，生成新的
+            life_attitude = generate_life_attitude(
+                age=age_info["age"],
+                gender=gender,
+                occupation=career_info["status"],
+                location=location,
+                values_orientation=personal_values.get("values_orientation", "")
+            )
+            print(f"✗ Life attitude not provided, generated based on user inputs")
+    else:
+        # 用户没提供，基于已有信息生成
+        life_attitude = generate_life_attitude(
+            age=age_info["age"],
+            gender=gender,
+            occupation=career_info["status"],
+            location=location,
+            values_orientation=personal_values.get("values_orientation", "")
+        )
+        print(f"✗ Life attitude not provided, generated based on user inputs")
+    
+    # 提取个人故事 - 如果用户提供了就用，否则生成
+    personal_story_data = web_profile.get("personal_story", {})
+    if isinstance(personal_story_data, dict) and personal_story_data.get("personal_story"):
+        story_text = personal_story_data.get("personal_story", "")
+        if story_text and story_text.strip():
+            # 用户提供了故事，直接使用
+            personal_story = {"personal_story": story_text.strip()}
+            print(f"✓ Using user-provided life story: {story_text[:50]}...")
+        else:
+            # 故事为空，生成新的
+            personal_story = generate_personal_story(
+                age=age_info["age"],
+                gender=gender,
+                occupation=career_info["status"],
+                location=location,
+                values_orientation=personal_values.get("values_orientation", ""),
+                life_attitude=life_attitude
+            )
+            print(f"✗ Life story not provided, generated based on user inputs")
+    elif isinstance(personal_story_data, str) and personal_story_data.strip():
+        # 用户提供了字符串格式的故事
+        personal_story = {"personal_story": personal_story_data.strip()}
+        print(f"✓ Using user-provided life story (string format)")
+    else:
+        # 用户没提供，基于已有信息生成
+        personal_story = generate_personal_story(
+            age=age_info["age"],
+            gender=gender,
+            occupation=career_info["status"],
+            location=location,
+            values_orientation=personal_values.get("values_orientation", ""),
+            life_attitude=life_attitude
+        )
+        print(f"✗ Life story not provided, generated based on user inputs")
+    
+    # 提取兴趣爱好 - 如果用户提供了就用，否则生成
+    interests = web_profile.get("interests", {})
+    interests_list = interests.get("interests", [])
+    if interests_list and len(interests_list) > 0 and any(i.strip() for i in interests_list):
+        # 用户提供了兴趣爱好，直接使用
+        interests = {"interests": [i.strip() for i in interests_list if i.strip()]}
+        print(f"✓ Using user-provided interests: {', '.join(interests['interests'])}")
+    else:
+        # 用户没提供，基于故事生成
+        interests = generate_interests_and_hobbies(personal_story)
+        print(f"✗ Interests not provided, generated based on life story")
+    
+    # 返回转换后的格式
+    return {
+        "age_info": age_info,
+        "gender": gender,
+        "location": location,
+        "career_info": career_info,
+        "personal_values": personal_values,
+        "life_attitude": life_attitude,
+        "personal_story": personal_story,
+        "interests": interests
+    }
+
+
 def generate_user_profile() -> Dict:
     """生成用户基础信息配置文件"""
     # 生成并存储直接函数返回值

generate_user_profile_final/code/web_api_bridge.py

@@ -103,7 +103,8 @@ def generate_profile_from_input(input_data: Dict[str, Any], attribute_count: int
             values_orientation = custom_personal_values.strip()
         else:
             print("Generating personal values...")
-            values_orientation = generate_personal_values(age, gender, occupation, location_info)
+            values_dict = generate_personal_values(age, gender, occupation, location_info)
+            values_orientation = values_dict.get("values_orientation", "") if isinstance(values_dict, dict) else str(values_dict)
         
         # Use custom life attitude if provided, otherwise generate it
         custom_life_attitude = custom_values.get('life_attitude')
@@ -148,9 +149,7 @@ def generate_profile_from_input(input_data: Dict[str, Any], attribute_count: int
                 "values_orientation": values_orientation
             },
             "life_attitude": life_attitude.get("outlook") if isinstance(life_attitude, dict) else str(life_attitude),
-            "personal_story": {
-                "personal_story": personal_story
-            },
+            "personal_story": personal_story if isinstance(personal_story, dict) else {"personal_story": personal_story},
             "interests": interests_and_hobbies
         }