diff --git a/dingo/model/llm/hhh/llm_text_3h.py b/dingo/model/llm/hhh/llm_text_3h.py
index 919d6bca..1887cc59 100644
--- a/dingo/model/llm/hhh/llm_text_3h.py
+++ b/dingo/model/llm/hhh/llm_text_3h.py
@@ -13,7 +13,7 @@ class LLMText3H(BaseOpenAI):
     def build_messages(cls, input_data):
         question = input_data.prompt
         response = input_data.content
-        prompt_content = cls.prompt.content % (question, response)
+        prompt_content = cls.prompt % (question, response)
 
         messages = [{"role": "user", "content": prompt_content}]
 
@@ -38,15 +38,21 @@ def process_response(cls, response: str) -> EvalDetail:
 
         result = EvalDetail(metric=cls.__name__)
 
+        # Get the quality dimension name from class name
+        # e.g., LLMText3HHelpful -> HELPFUL
+        class_prefix = "LLMText3H"
+        if cls.__name__.startswith(class_prefix):
+            quality_name = cls.__name__[len(class_prefix):].upper()
+        else:
+            quality_name = cls.__name__.upper()
+
         # eval_status
         if response_model.score == 1:
-            tmp_name = cls.prompt.__name__[8:].upper()
-            result.label = [f"{QualityLabel.QUALITY_GOOD}.{tmp_name}"]
+            result.label = [f"{QualityLabel.QUALITY_GOOD}.{quality_name}"]
             result.reason = [response_model.reason] if response_model.reason else ["Response meets quality criteria"]
         else:
             result.status = True
-            tmp_name = "NOT_" + cls.prompt.__name__[8:].upper()
-            result.label = [f"QUALITY_BAD.{tmp_name}"]
+            result.label = [f"QUALITY_BAD.NOT_{quality_name}"]
             result.reason = [response_model.reason] if response_model.reason else ["Response fails quality criteria"]
 
         return result
diff --git a/docs/artimuse.md b/docs/artimuse.md
index f71b6191..f5d92e23 100644
--- a/docs/artimuse.md
+++ b/docs/artimuse.md
@@ -24,7 +24,7 @@ RuleImageArtimuse 基于 ArtiMuse 在线服务对输入图片进行美学质量
 
 ## 核心方法
 
-### `eval(cls, input_data: Data) -> ModelRes`
+### `eval(cls, input_data: Data) -> EvalDetail`
 
 这是规则的主要评估方法，接收包含图像 URL 的 `Data` 对象，返回评估结果。
 
@@ -50,20 +50,19 @@ RuleImageArtimuse 基于 ArtiMuse 在线服务对输入图片进行美学质量
 
 #### 返回值
 
-返回 `ModelRes` 对象，包含以下属性：
+返回 `EvalDetail` 对象，包含以下属性：
 
-- `eval_status`: 布尔值，表示图像质量是否不合格（低于阈值）
-- `type`: 评估结果类型（"Artimuse_Succeeded" 或 "Artimuse_Fail"）
-- `name`: 评估结果名称（"BadImage" 或 "GoodImage" 或 "Exception"）
+- `metric`: 指标名称（"RuleImageArtimuse"）
+- `status`: 布尔值，表示图像质量是否不合格（低于阈值）(True=不合格, False=合格)
+- `label`: 质量标签列表（如 ["Artimuse_Succeeded.BadImage"] 或 ["QUALITY_GOOD"]）
 - `reason`: 包含详细评估信息或异常信息的数组（字符串化 JSON）
 
 ## 异常处理
 
-当评估过程中发生异常时，返回的 `ModelRes` 对象将包含：
+当评估过程中发生异常时，返回的 `EvalDetail` 对象将包含：
 
-- `eval_status`: `False`
-- `type`: `"Artimuse_Fail"`
-- `name`: `"Exception"`
+- `status`: `False`
+- `label`: `["Artimuse_Fail.Exception"]`
 - `reason`: 包含异常信息的数组
 
 ## 使用示例
diff --git a/docs/ats_resume_guide.md b/docs/ats_resume_guide.md
index f137c36a..a3254230 100644
--- a/docs/ats_resume_guide.md
+++ b/docs/ats_resume_guide.md
@@ -16,6 +16,12 @@ ATS 工具套件用于：
 
 分析简历与 JD 的匹配度，输出加权匹配分数和详细分析报告。
 
+**核心功能：**
+- 语义匹配（不仅是字符串匹配）
+- 同义词自动识别（如 k8s → Kubernetes）
+- 负向约束识别（Excluded 技能警告）
+- 基于证据的匹配（引用简历原文）
+
 **输入字段：**
 | 字段 | 类型 | 必需 | 说明 |
 |------|------|------|------|
@@ -26,8 +32,17 @@ ATS 工具套件用于：
 | 字段 | 类型 | 说明 |
 |------|------|------|
 | `score` | float | 匹配分数 (0.0-1.0) |
-| `error_status` | bool | 是否低于阈值 (默认 0.6) |
-| `reason` | List[str] | 详细分析报告 |
+| `status` | bool | 是否低于阈值 (True=低于，False=通过) |
+| `reason` | List[str] | 详细分析报告（文本格式） |
+
+**内置同义词映射 (SYNONYM_MAP)：**
+```
+k8s → Kubernetes, js → JavaScript, ts → TypeScript
+py → Python, tf → TensorFlow, pt → PyTorch
+nodejs → Node.js, postgres → PostgreSQL
+aws → Amazon Web Services, gcp → Google Cloud Platform
+ml → Machine Learning, dl → Deep Learning, nlp → NLP
+```
 
 ### 2. LLMResumeOptimizer（简历优化器）
 
@@ -80,6 +95,8 @@ jd = """
 match_data = Data(data_id='test_1', content=resume, prompt=jd)
 match_result = LLMKeywordMatcher.eval(match_data)
 print(f"匹配分数: {match_result.score}")
+print(f"是否通过: {'通过' if not match_result.status else '未通过'}")
+print(f"分析报告: {match_result.reason[0]}")
 
 # Step 2: 简历优化
 optimize_data = Data(
@@ -89,18 +106,23 @@ optimize_data = Data(
     context='{"match_details": {"missing": [{"skill": "Docker", "importance": "Required"}]}}'
 )
 opt_result = LLMResumeOptimizer.eval(optimize_data)
-print(f"优化结果: {opt_result.reason[0]}")
+print(f"优化摘要: {opt_result.reason[0]}")
+print(f"完整结果: {opt_result.optimized_content}")
 ```
 
 ## 📊 匹配分数计算
 
-### 权重分配
+### 权重公式
+
+```
+score = (Required_Matched × 2 + Nice_Matched × 1) / (Required_Total × 2 + Nice_Total × 1)
+```
 
 | 类别 | 权重 | 说明 |
 |------|------|------|
-| Required (必需) | 0.7 | 缺失会显著降低分数 |
-| Nice-to-have (加分) | 0.3 | 缺失影响较小 |
-| Excluded (排除) | -0.1 | 存在会扣分 |
+| Required (必需) | ×2 | 缺失会显著降低分数 |
+| Nice-to-have (加分) | ×1 | 缺失影响较小 |
+| Excluded (排除) | 不计分 | 仅生成警告，不影响分数 |
 
 ### 阈值配置
 
@@ -156,16 +178,31 @@ Nice-to-have (Missing): Kubernetes
 
 ### ResumeOptimizer 输出
 
-结果同样存放在 `result.reason[0]` 中，JSON 格式：
+**`reason[0]`**: 人类可读的摘要文本
+**`optimized_content`**: 完整的 JSON 优化结果
 
 ```python
 # 访问方式
 result = LLMResumeOptimizer.eval(data)
-import json
-output = json.loads(result.reason[0])
+
+# 摘要文本
+print(result.reason[0])
+
+# 完整 JSON 结果
+opt = result.optimized_content
+print(opt.get('optimization_summary'))
+print(opt.get('section_changes'))
 ```
 
 **`reason[0]` 内容示例：**
+```
+Overall: 优化了专业技能板块
+Keywords Added: Docker
+Associative: Kubernetes (了解概念)
+Sections Modified: 专业技能
+```
+
+**`optimized_content` 结构：**
 ```json
 {
   "optimization_summary": {
diff --git a/docs/document_ocr.md b/docs/document_ocr.md
index 30f7176e..9f0206e8 100644
--- a/docs/document_ocr.md
+++ b/docs/document_ocr.md
@@ -22,9 +22,7 @@ Dingo 提供了一种基于LLM的文档OCR解析质量评估工具，可帮助
 dingo/
   ├── model/
   │   ├── llm/
-  │   │   └── vlm_document_parsing.py         # 评估器实现
-  │   └── prompt/
-  │       └── prompt_mineru_recognize.py      # 评估提示词
+  │   │   └── llm_document_parsing_ocr.py     # 评估器实现（含内嵌Prompt）
   │── examples/
   │   └── document_parser/
   │       └── document_parsing_quality_ocr.py  # 单条评估示例
@@ -75,11 +73,11 @@ input_data = {
 #### 输出结果格式
 
 ```python
-# result 是 ModelRes 对象，包含以下字段：
-result.type          # 错误问题一级标签: prompt中定义的一级错误大类
-result.name          # 错误问题二级标签: 一级错误大类对应的详细错误标签 List[str]
-result.eval_status  # 错误状态: False 或 True
-result.reason        # 评估原因: List[str]
+# result 是 EvalDetail 对象，包含以下字段：
+result.metric        # 指标名称: "LLMMinerURecognizeQuality"
+result.label         # 错误标签列表: ["error_category1.error_category2.error_label1.error_label2"]
+result.status        # 错误状态: False (默认值)
+result.reason        # 评估原因: List[str]，包含完整的JSON分析结果
 ```
 
 
diff --git a/docs/document_parsing_quality_guide.md b/docs/document_parsing_quality_guide.md
index 9732074d..e9ca048b 100644
--- a/docs/document_parsing_quality_guide.md
+++ b/docs/document_parsing_quality_guide.md
@@ -1,4 +1,4 @@
-# VLMDocumentParsingQuality 文档解析评估工具 使用文档
+# VLMDocumentParsing 文档解析评估工具 使用文档
 
 Dingo 提供了一种基于VLM的文档解析质量评估与可视化工具，可帮助您：
 - 评估文档解析模型输出质量
@@ -6,7 +6,7 @@ Dingo 提供了一种基于VLM的文档解析质量评估与可视化工具，
 
 ## 工具介绍
 
-### VLMDocumentParsingQuality：文档解析评估工具
+### VLMDocumentParsing：文档解析评估工具
 
 #### 功能说明
 该工具用于评估文档解析模型效果，具体功能包括：
@@ -22,9 +22,8 @@ Dingo 提供了一种基于VLM的文档解析质量评估与可视化工具，
 dingo/
   ├── model/
   │   ├── llm/
-  │   │   └── vlm_document_parsing.py         # 评估器实现
-  │   └── prompt/
-  │       └── prompt_document_parsing.py      # 评估提示词
+  │   │   └── mineru/
+  │   │       └── vlm_document_parsing.py     # 评估器实现（含内嵌Prompt）
   │── examples/
   │   └── document_parser/
   │       └── vlm_document_parser_quality.py  # 单条评估示例
@@ -64,7 +63,7 @@ input_data = {
         },
         "evaluator": {
             "llm_config": {
-                "VLMDocumentParsingQuality": {
+                "VLMDocumentParsing": {
                     "key": "",
                     "api_url": "",
                 }
@@ -76,11 +75,11 @@ input_data = {
 #### 输出结果格式
 
 ```python
-# result 是 ModelRes 对象，包含以下字段：
-result.type          # 错误问题一级标签: prompt中定义的一级错误大类
-result.name          # 错误问题二级标签: 一级错误大类对应的详细错误标签 List[str]
-result.eval_status  # 错误状态: False 或 True
-result.reason        # 评估原因: List[str]
+# result 是 EvalDetail 对象，包含以下字段：
+result.metric        # 指标名称: "VLMDocumentParsing"
+result.label         # 错误标签列表: ["公式相关问题.行内公式漏检", "表格相关问题.单元格内容错误"]
+result.status        # 错误状态: False (默认值，该类不设置)
+result.reason        # 评估原因: List[str]，包含完整的JSON分析结果
 ```
 
 
@@ -114,7 +113,7 @@ if __name__ == '__main__':
         },
         "evaluator": {
             "llm_config": {
-                "VLMDocumentParsingQuality": {
+                "VLMDocumentParsing": {
                     "key": "",
                     "api_url": "",
                 }
diff --git a/docs/factcheck_guide.md b/docs/factcheck_guide.md
index c11d52f4..4112707f 100644
--- a/docs/factcheck_guide.md
+++ b/docs/factcheck_guide.md
@@ -64,14 +64,10 @@ data = Data(
 # 执行评估
 result = LLMFactCheckPublic.eval(data)
 
-# 查看结果
-print(f"Factual ratio: {result.score:.2%}")
-print(f"Reason: {result.reason}")
-print("\nDetailed results:")
-for claim in result.raw_resp["results"]:
-    print(f"\nClaim: {claim.claim}")
-    print(f"Answer: {claim.answer}")
-    print(f"Reasoning: {claim.reasoning}")
+# 查看结果 (返回 EvalDetail 对象)
+print(f"是否通过: {'通过' if not result.status else '未通过'}")
+print(f"标签: {result.label}")
+print(f"详细原因: {result.reason[0]}")
 ```
 
 ### 场景二：评估数据集
@@ -143,13 +139,10 @@ rag_data = {
 data = Data(**rag_data)
 result = LLMFactCheckPublic.eval(data)
 
-# 分析结果
-print(f"Factual consistency: {result.score:.2%}")
-for claim in result.raw_resp["results"]:
-    if claim.answer != "true":
-        print(f"\nPotential hallucination:")
-        print(f"Claim: {claim.claim}")
-        print(f"Evidence: {claim.reasoning}")
+# 分析结果 (返回 EvalDetail 对象)
+print(f"是否通过: {'通过' if not result.status else '未通过'}")
+print(f"标签: {result.label}")
+print(f"详细原因: {result.reason[0]}")
 ```
 
 ### 场景四：多轮对话监控
@@ -173,9 +166,10 @@ for turn in conversation:
     data = Data(**turn)
     result = LLMFactCheckPublic.eval(data)
     print(f"\nTurn {turn['data_id']}:")
-    print(f"Factual ratio: {result.score:.2%}")
-    if result.score < LLMFactCheckPublic.threshold:
+    print(f"是否通过: {'通过' if not result.status else '未通过'}")
+    if result.status:
         print("Warning: Potential misinformation detected!")
+        print(f"详情: {result.reason[0]}")
 ```
 
 ## 最佳实践
@@ -241,30 +235,16 @@ dingo/
 ### 评估结果格式
 
 ```python
-ModelRes(
-    score=0.85,  # 事实性得分
-    threshold=0.8,  # 判断阈值
-    reason=["Found 10 claims: 8 true, 1 false, 1 unsure..."],
-    raw_resp={
-        "claims": ["claim1", "claim2", ...],
-        "results": [
-            FactCheckResult(
-                claim="...",
-                answer="true",
-                reasoning="...",
-                supporting_evidence=[...]
-            ),
-            ...
-        ],
-        "metrics": {
-            "factual_ratio": 0.85,
-            "true_count": 8,
-            "false_count": 1,
-            "unsure_count": 1,
-            "total_claims": 10
-        }
-    }
+# LLMFactCheckPublic 返回 EvalDetail 对象
+EvalDetail(
+    metric="LLMFactCheckPublic",           # 指标名称
+    status=False,                           # 是否未通过 (False=通过, True=未通过)
+    label=["QUALITY_GOOD.FACTUALITY_CHECK_PASSED"],  # 质量标签
+    reason=["Found 10 claims: 8 true, 1 false, 1 unsure. Factual ratio: 80.00%"]
 )
+
+# reason[0] 包含完整的评估摘要，格式示例：
+# "Found 10 claims: 8 true, 1 false, 1 unsure. Factual ratio: 80.00%"
 ```
 
 ## 参考资料
diff --git a/docs/hallucination_guide.md b/docs/hallucination_guide.md
index 50ecb9a6..2ca58899 100644
--- a/docs/hallucination_guide.md
+++ b/docs/hallucination_guide.md
@@ -88,8 +88,7 @@ data = Data(
 result = RuleHallucinationHHEM.eval(data)
 
 # 查看结果
-print(f"是否检测到幻觉: {result.eval_status}")
-print(f"HHEM 分数: {getattr(result, 'score', 'N/A')}")
+print(f"是否检测到幻觉: {result.status}")  # True=检测到幻觉, False=未检测到
 print(f"详细分析: {result.reason[0]}")
 ```
 
@@ -122,9 +121,8 @@ data = Data(
 result = LLMHallucination.eval(data)
 
 # 查看结果
-print(f"是否检测到幻觉: {result.eval_status}")
-print(f"幻觉分数: {getattr(result, 'score', 'N/A')}")
-print(f"详细原因: {result.reason[0]}")
+print(f"是否检测到幻觉: {result.status}")  # True=检测到幻觉, False=未检测到
+print(f"详细原因: {result.reason[0]}")  # 包含幻觉分数等详细信息
 ```
 
 ## 📊 批量数据集评估
@@ -280,21 +278,30 @@ results = RuleHallucinationHHEM.batch_evaluate(data_list)  # 批量更高效
 
 ## 📊 输出结果解析
 
-### ModelRes 字段说明
+### RuleHallucinationHHEM (EvalDetail) 字段说明
 
 ```python
-result = RuleHallucinationHHEM.eval(data)  # 或 LLMHallucination.eval(data)
+result = RuleHallucinationHHEM.eval(data)
 
-# 标准字段
-result.eval_status      # bool: 是否检测到幻觉
-result.type             # str: 质量类型标识
-result.name             # str: 检测结果名称
+# 标准字段 (EvalDetail)
+result.metric           # str: 指标名称 ("RuleHallucinationHHEM")
+result.status           # bool: 是否检测到幻觉 (True=有幻觉, False=无幻觉)
+result.label            # List[str]: 质量标签 (如 ["QUALITY_BAD_HALLUCINATION.HALLUCINATION_DETECTED"])
 result.reason           # List[str]: 详细分析原因
+```
+
+### LLMHallucination (EvalDetail) 字段说明
+
+```python
+result = LLMHallucination.eval(data)
 
-# 扩展字段
-result.score            # float: 幻觉分数 (0.0-1.0)
-result.verdict_details  # List[str]: 每个上下文的判断详情（GPT 模式）
-result.consistency_scores # List[float]: HHEM 原始一致性分数（HHEM 模式）
+# 标准字段 (EvalDetail)
+result.metric           # str: 指标名称 ("LLMHallucination")
+result.status           # bool: 是否检测到幻觉 (True=有幻觉, False=无幻觉)
+result.label            # List[str]: 质量标签
+                        # 有幻觉: ["QUALITY_BAD_HALLUCINATION.HALLUCINATION_DETECTED"]
+                        # 无幻觉: ["QUALITY_GOOD.NO_HALLUCINATION"]
+result.reason           # List[str]: 详细分析原因（包含幻觉分数信息）
 ```
 
 ### 典型输出示例
@@ -357,7 +364,7 @@ def monitor_rag_response(question, generated_answer, retrieved_docs):
 
     result = RuleHallucinationHHEM.eval(data)  # 本地、快速、免费
 
-    if result.eval_status:
+    if result.status:
         logger.warning(f"检测到幻觉: {result.reason[0]}")
         # 触发人工审核或回答重生成
 ```
@@ -387,7 +394,7 @@ def filter_hallucinated_responses(responses_with_context):
         # 使用本地HHEM进行快速检测
         result = RuleHallucinationHHEM.eval(data)
 
-        if not result.eval_status:  # 无幻觉
+        if not result.status:  # 无幻觉
             clean_responses.append(item)
         else:
             log_quality_issue(item, result.reason[0])
@@ -427,7 +434,7 @@ class RAGWithHallucinationDetection:
         hallucination_result = self.detector.eval(data)
 
         # 4. 根据检测结果决定是否返回答案
-        if hallucination_result.eval_status:
+        if hallucination_result.status:
             self.log_hallucination(question, generated_answer, hallucination_result)
             return {
                 "answer": None,
diff --git a/docs/html_extract_compare_v2.md b/docs/html_extract_compare_v2.md
index 6637fae2..c0d92242 100644
--- a/docs/html_extract_compare_v2.md
+++ b/docs/html_extract_compare_v2.md
@@ -81,20 +81,20 @@ data = Data(
 ## 输出结果格式
 
 ```python
-# result 是 ModelRes 对象，包含以下字段：
-result.type          # 判断类型: "TOOL_ONE_BETTER" / "TOOL_EQUAL" / "TOOL_TWO_BETTER"
-result.name          # 判断名称: "Judgement_A" / "Judgement_B" / "Judgement_C"
-result.eval_status  # 错误状态: False (A/B) 或 True (C)
+# result 是 EvalDetail 对象，包含以下字段：
+result.metric        # 指标名称: "LLMHtmlExtractCompareV2"
+result.label         # 判断标签: ["TOOL_ONE_BETTER.Judgement_A"] 等
+result.status        # 错误状态: False (A/B) 或 True (C)
 result.reason        # 推理过程: List[str]
 ```
 
 ### 结果映射
 
-| 判断结果 | `result.type` | `result.name` | `result.eval_status` | 含义 |
-|----------|---------------|---------------|----------------------|------|
-| A | TOOL_ONE_BETTER | Judgement_A | False | 工具A提取的信息更完整 |
-| B | TOOL_EQUAL | Judgement_B | False | 两个工具提取的信息量相同 |
-| C | TOOL_TWO_BETTER | Judgement_C | True | 工具B提取的信息更完整 |
+| 判断结果 | `result.label` | `result.status` | 含义 |
+|----------|----------------|-----------------|------|
+| A | ["TOOL_ONE_BETTER.Judgement_A"] | False | 工具A提取的信息更完整 |
+| B | ["TOOL_EQUAL.Judgement_B"] | False | 两个工具提取的信息量相同 |
+| C | ["TOOL_TWO_BETTER.Judgement_C"] | True | 工具B提取的信息更完整 |
 
 ## 使用示例
 
@@ -125,7 +125,7 @@ data = Data(
 result = evaluator.eval(data)
 
 # 查看结果
-print(f"判断: {result.type}")
+print(f"判断: {result.label}")
 print(f"推理: {result.reason[0]}")
 ```
 
diff --git a/docs/image_lable_check_guide.md b/docs/image_lable_check_guide.md
index cfa3fbd2..7b3818f0 100644
--- a/docs/image_lable_check_guide.md
+++ b/docs/image_lable_check_guide.md
@@ -235,39 +235,28 @@ if __name__ == '__main__':
 #### RuleImageLabelOverlap 输出结果格式：
 
 ```python
-ModelRes(
-    name="RuleImageLabelOverlap" or "GOOD_IMG_LABEL",
-    type="IMG_LABEL_OVERLAP" or "NO_LABEL_OVERLAP",
-    eval_status=True/False,  # 是否存在符合阈值的重叠
-    reason=[json.dumps({
-        "id": data_id,
-        "has_overlap": True/False,
-        "overlap_stats": {
-            "full_overlap_pairs": 完全重叠框数量,
-            "partial_overlap_pairs": 部分重叠框数量,
-            "total_boxes": 总边界框数
-        },
-        "visualization_path": 图像保存路径
-    })]
+EvalDetail(
+    metric="RuleImageLabelOverlap",
+    status=True/False,  # 是否存在符合阈值的重叠
+    label=["LabelOverlap_Fail.RuleImageLabelOverlap"],  # 存在重叠时设置
+    reason=["重叠检测：完全重叠=N，部分重叠=M"]  # 重叠统计信息
 )
 ```
 
 #### RuleImageLabelVisualization 输出结果格式：
 ```python
-ModelRes(
-    name="RuleImageLabelVisualization" or "NO_LABEL_DATA",
-    type="IMG_LABEL_VISUALIZATION" or "NO_IMG_LABEL_VISUALIZATION",
-    eval_status=True/False,  # 是否发生错误
-    reason=[json.dumps({
-        "id": data_id,
-        "visualization_status": "success",
-        "original_image_path": 原始图像路径,
-        "visualization_path": 可视化图像路径,
-        "label_stats": {
-            "total_labels": 总标注数,
-            "top_level_labels": 顶层标注数
-        }
-    })]
+EvalDetail(
+    metric="RuleImageLabelVisualization",
+    status=False,  # 成功时为False
+    label=None,    # 成功时不设置label
+    reason=None    # 成功时不设置reason
+)
+# 错误时：
+EvalDetail(
+    metric="RuleImageLabelVisualization",
+    status=False,
+    label=["LabelVisualization_Fail.错误类型"],  # 如ParseError, InvalidAnnotationType等
+    reason=["错误描述信息"]
 )
 ```
 
diff --git a/docs/image_quality_check_guide.md b/docs/image_quality_check_guide.md
index 2521b0ab..9c096455 100644
--- a/docs/image_quality_check_guide.md
+++ b/docs/image_quality_check_guide.md
@@ -432,54 +432,64 @@ if __name__ == '__main__':
 
 ### 10.2 输出结果格式
 
-#### RuleImageValid 输出结果格式：
+所有图像规则返回 `EvalDetail` 对象，包含以下字段：
 
 ```python
-ModelRes(
-    name="RuleImageValid",
-    type="QUALITY_BAD_IMG_EFFECTIVENESS",
-    eval_status=True/False,  # 是否为无效图像
-    reason=["Image is not valid: all white or black"]  # 错误原因
+EvalDetail(
+    metric="RuleImageValid",    # 指标名称
+    status=True/False,          # 是否未通过 (True=未通过, False=通过)
+    label=["QUALITY_BAD_IMG_EFFECTIVENESS.RuleImageValid"],  # 质量标签
+    reason=["Image is not valid: all white or black"]  # 详细原因
 )
 ```
 
-#### RuleImageSizeValid 输出结果格式：
+#### RuleImageValid 输出结果示例：
 ```python
-ModelRes(
-    name="RuleImageSizeValid",
-    type="QUALITY_BAD_IMG_EFFECTIVENESS",
-    eval_status=True/False,  # 图像尺寸是否无效
-    reason=["Image size is not valid, the ratio of width to height: 比值"]  # 错误原因
+EvalDetail(
+    metric="RuleImageValid",
+    status=True,  # 是否为无效图像
+    label=["QUALITY_BAD_IMG_EFFECTIVENESS.RuleImageValid"],
+    reason=["Image is not valid: all white or black"]
 )
 ```
 
-#### RuleImageQuality 输出结果格式：
+#### RuleImageSizeValid 输出结果示例：
 ```python
-ModelRes(
-    name="RuleImageQuality",
-    type="QUALITY_BAD_IMG_EFFECTIVENESS",
-    eval_status=True/False,  # 图像质量是否不满足要求
-    reason=["Image quality is not satisfied, ratio: 评分值"]  # 错误原因
+EvalDetail(
+    metric="RuleImageSizeValid",
+    status=True,  # 图像尺寸是否无效
+    label=["QUALITY_BAD_IMG_EFFECTIVENESS.RuleImageSizeValid"],
+    reason=["Image size is not valid, the ratio of width to height: 比值"]
 )
 ```
 
-#### RuleImageRepeat 输出结果格式：
+#### RuleImageQuality 输出结果示例：
 ```python
-ModelRes(
-    name="RuleImageRepeat",
-    type="QUALITY_BAD_IMG_SIMILARITY",
-    eval_status=True/False,  # 是否存在重复图像
+EvalDetail(
+    metric="RuleImageQuality",
+    status=True,  # 图像质量是否不满足要求
+    label=["QUALITY_BAD_IMG_EFFECTIVENESS.RuleImageQuality"],
+    reason=["Image quality is not satisfied, ratio: 评分值"]
+)
+```
+
+#### RuleImageRepeat 输出结果示例：
+```python
+EvalDetail(
+    metric="RuleImageRepeat",
+    status=True,  # 是否存在重复图像
+    label=["QUALITY_BAD_IMG_SIMILARITY.RuleImageRepeat"],
     reason=["图像1 -> [重复图像列表]", ..., {"duplicate_ratio": 重复率}]
 )
 ```
 
-#### RuleImageTextSimilarity 输出结果格式：
+#### RuleImageTextSimilarity 输出结果示例：
 ```python
-ModelRes(
-    name="RuleImageTextSimilarity",
-    type="QUALITY_BAD_IMG_RELEVANCE",
-    eval_status=True/False,  # 图像与文本相似度是否不足
-    reason=["Image quality is not satisfied, ratio: 相似度值"]  # 错误原因
+EvalDetail(
+    metric="RuleImageTextSimilarity",
+    status=True,  # 图像与文本相似度是否不足
+    label=["QUALITY_BAD_IMG_RELEVANCE.RuleImageTextSimilarity"],
+    reason=["Image quality is not satisfied, ratio: 相似度值"]
 )
 ```
 
diff --git a/docs/layout_quality_guide.md b/docs/layout_quality_guide.md
index a28b3dbd..3210b5b5 100644
--- a/docs/layout_quality_guide.md
+++ b/docs/layout_quality_guide.md
@@ -21,9 +21,7 @@ Dingo 提供了一种基于VLM的Layout布局检测质量评估，可帮助您
 dingo/
   ├── model/
   │   ├── llm/
-  │   │   └── vlm_layout_quality.py         # 评估器实现
-  │   └── prompt/
-  │       └── prompt_layout_quality.py      # 评估提示词
+  │   │   └── vlm_layout_quality.py         # 评估器实现（含内嵌Prompt）
   │── examples/
   │   └── document_parser/
   │       └── vlm_layout_quality.py         # 评估示例
@@ -36,7 +34,7 @@ dingo/
 ```
 
 ##### 评估提示词
-我们的评估效果依赖于精心设计的 Prompt。其核心思想是：
+我们的评估效果依赖于精心设计的 Prompt（内嵌在 `vlm_layout_quality.py` 中）。其核心思想是：
 
 1. Layout布局检测元素列别，我们基于Mineru的输出类型，来设定提示词。
 2. 分层错误标签：我们将布局检测问题分为5个大类：检测遗漏错误、检测不准错误、类别错误、阅读顺序错、其他错误。
@@ -79,11 +77,11 @@ input_data = {
 #### 输出结果格式
 
 ```python
-# result 是 ModelRes 对象，包含以下字段：
-result.type          # 错误问题一级标签: prompt中定义错误类别
-result.name          # 错误描述: 错误列别对应的详细错描述
-result.eval_status  # 错误状态: False 或 True
-result.reason        # 评估原因: List[str]
+# result 是 EvalDetail 对象，包含以下字段：
+result.metric        # 指标名称: "VLMLayoutQuality"
+result.label         # 错误标签列表: 从JSON响应中提取的eval_details字段列表
+result.status        # 错误状态: False (默认值，该类不设置)
+result.reason        # 评估原因: List[str]，包含完整的JSON分析结果
 ```
 
 
diff --git a/examples/3h/3h_eval.py b/examples/3h/3h_eval.py
index a0c7b0fe..08d78941 100644
--- a/examples/3h/3h_eval.py
+++ b/examples/3h/3h_eval.py
@@ -4,10 +4,14 @@
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
-    OPENAI_MODEL = 'deepseek-chat'
-    OPENAI_URL = 'https://api.deepseek.com/v1'
-    OPENAI_KEY = os.getenv("OPENAI_KEY")
+    # Configure LLM (set your API key via environment variable OPENAI_KEY)
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "gpt-4o")
+    OPENAI_URL = os.getenv("OPENAI_URL", "https://api.openai.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_KEY", "YOUR_API_KEY")  # Set OPENAI_KEY env var
     common_config = {
         "model": OPENAI_MODEL,
         "key": OPENAI_KEY,
@@ -15,7 +19,7 @@
     }
 
     input_data = {
-        "input_path": str(Path("test/data/test_3h_jsonl.jsonl")),
+        "input_path": str(PROJECT_ROOT / "test/data/test_3h_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl"
diff --git a/examples/artimuse/artimuse.py b/examples/artimuse/artimuse.py
index c8ee0d91..673d09ab 100644
--- a/examples/artimuse/artimuse.py
+++ b/examples/artimuse/artimuse.py
@@ -1,9 +1,14 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_imgae_artimuse.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_imgae_artimuse.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl"
diff --git a/examples/ats_resume/sdk_keyword_matcher.py b/examples/ats_resume/sdk_keyword_matcher.py
index a2d9be42..29132cb7 100644
--- a/examples/ats_resume/sdk_keyword_matcher.py
+++ b/examples/ats_resume/sdk_keyword_matcher.py
@@ -20,15 +20,17 @@
 - Missing required/nice-to-have skills
 """
 
+import os
+
 from dingo.config.input_args import EvaluatorLLMArgs
 from dingo.io.input import Data
 from dingo.model.llm.llm_keyword_matcher import LLMKeywordMatcher
 
-# Configure LLM
+# Configure LLM (set your API key via environment variable OPENAI_KEY)
 LLMKeywordMatcher.dynamic_config = EvaluatorLLMArgs(
-    key='sk-xxx',  # Replace with your API key
-    api_url='https://api.deepseek.com',
-    model='deepseek-chat',
+    key=os.getenv("OPENAI_KEY", "YOUR_API_KEY"),  # Replace with your API key or set OPENAI_KEY env var
+    api_url=os.getenv("OPENAI_URL", "https://api.openai.com/v1"),
+    model=os.getenv("OPENAI_MODEL", "gpt-4o"),
 )
 
 
@@ -74,7 +76,7 @@ def example_1_basic_matching():
     result = LLMKeywordMatcher.eval(data)
 
     print(f"Match Score: {getattr(result, 'score', 'N/A')}")
-    print(f"Error Status: {result.error_status}")
+    print(f"Status: {result.status}")  # True = has issues, False = passed
     print(f"Reason:\n{result.reason[0]}")
     print()
 
@@ -121,7 +123,7 @@ def example_2_english_resume():
     result = LLMKeywordMatcher.eval(data)
 
     print(f"Match Score: {getattr(result, 'score', 'N/A')}")
-    print(f"Error Status: {result.error_status}")
+    print(f"Status: {result.status}")  # True = has issues, False = passed
     print(f"Reason:\n{result.reason[0]}")
     print()
 
@@ -155,7 +157,7 @@ def example_3_low_match():
     result = LLMKeywordMatcher.eval(data)
 
     print(f"Match Score: {getattr(result, 'score', 'N/A')}")
-    print(f"Error Status: {result.error_status}")  # Should be True (low match)
+    print(f"Status: {result.status}")  # True = has issues (low match), False = passed
     print(f"Reason:\n{result.reason[0]}")
     print()
 
diff --git a/examples/ats_resume/sdk_resume_optimizer.py b/examples/ats_resume/sdk_resume_optimizer.py
index 53fbf6a6..a3f78d5a 100644
--- a/examples/ats_resume/sdk_resume_optimizer.py
+++ b/examples/ats_resume/sdk_resume_optimizer.py
@@ -18,15 +18,17 @@
 - Section-by-section changes
 """
 
+import os
+
 from dingo.config.input_args import EvaluatorLLMArgs
 from dingo.io.input import Data
 from dingo.model.llm.llm_resume_optimizer import LLMResumeOptimizer
 
-# Configure LLM
+# Configure LLM (从环境变量读取)
 LLMResumeOptimizer.dynamic_config = EvaluatorLLMArgs(
-    key='sk-xxx',  # Replace with your API key
-    api_url='https://api.deepseek.com',
-    model='deepseek-chat',
+    key=os.getenv("OPENAI_API_KEY", ""),
+    api_url=os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com"),
+    model=os.getenv("OPENAI_MODEL", "deepseek-chat"),
 )
 
 
diff --git a/examples/audio/audioSnr.py b/examples/audio/audioSnr.py
index 8be62d11..3544e663 100644
--- a/examples/audio/audioSnr.py
+++ b/examples/audio/audioSnr.py
@@ -1,12 +1,14 @@
-import os
 from pathlib import Path
 
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": str(Path("test/data/test_audio_snr.jsonl")),
+        "input_path": str(PROJECT_ROOT / "test/data/test_audio_snr.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/classify/sdk_QR_classification.py b/examples/classify/sdk_QR_classification.py
index 1c528f2b..1cd60330 100644
--- a/examples/classify/sdk_QR_classification.py
+++ b/examples/classify/sdk_QR_classification.py
@@ -1,10 +1,15 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def classify_QR():
     input_data = {
-        "input_path": "../../test/data/test_imgQR_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_imgQR_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/classify/sdk_topic_classifcation.py b/examples/classify/sdk_topic_classifcation.py
index b95591c0..2055a5f4 100644
--- a/examples/classify/sdk_topic_classifcation.py
+++ b/examples/classify/sdk_topic_classifcation.py
@@ -1,10 +1,23 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
+# Configure LLM (set your API key via environment variable OPENAI_KEY)
+LLM_CONFIG = {
+    "key": os.getenv("OPENAI_KEY", "YOUR_API_KEY"),
+    "api_url": os.getenv("OPENAI_URL", "https://api.openai.com/v1"),
+    "model": os.getenv("OPENAI_MODEL", "gpt-4o")
+}
+
 
 def classify_topic():
     input_data = {
-        "input_path": "../../test/data/test_sft_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_sft_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl"
@@ -19,7 +32,7 @@ def classify_topic():
             {
                 "fields": {"content": "question"},
                 "evals": [
-                    {"name": "LLMClassifyTopic", "config": {"key": "", "api_url": ""}}
+                    {"name": "LLMClassifyTopic", "config": LLM_CONFIG}
                 ]
             }
         ]
diff --git a/examples/compare/compare_code.py b/examples/compare/compare_code.py
index 14c21c97..942533c8 100644
--- a/examples/compare/compare_code.py
+++ b/examples/compare/compare_code.py
@@ -1,8 +1,13 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 input_data = {
-    'input_path': '../../test/data/compare/WebMainBench_test_1011_dataset_with_results_clean.jsonl',
+    'input_path': str(PROJECT_ROOT / 'test/data/compare/WebMainBench_test_1011_dataset_with_results_clean.jsonl'),
     'dataset': {
         'source': 'local',
         'format': 'jsonl',
diff --git a/examples/compare/compare_math.py b/examples/compare/compare_math.py
index b027fb4a..46f1fba6 100644
--- a/examples/compare/compare_math.py
+++ b/examples/compare/compare_math.py
@@ -1,8 +1,13 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 input_data = {
-    'input_path': '../../test/data/compare/WebMainBench_test_1011_dataset_with_results_clean.jsonl',
+    'input_path': str(PROJECT_ROOT / 'test/data/compare/WebMainBench_test_1011_dataset_with_results_clean.jsonl'),
     'dataset': {
         'source': 'local',
         'format': 'jsonl',
diff --git a/examples/compare/compare_table.py b/examples/compare/compare_table.py
index 9d9f2426..3b5bb3ce 100644
--- a/examples/compare/compare_table.py
+++ b/examples/compare/compare_table.py
@@ -1,8 +1,13 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 input_data = {
-    'input_path': '../../test/data/compare/WebMainBench_test_1011_dataset_with_results_clean_llm_webkit_html.jsonl',
+    'input_path': str(PROJECT_ROOT / 'test/data/compare/WebMainBench_test_1011_dataset_with_results_clean_llm_webkit_html.jsonl'),
     'dataset': {
         'source': 'local',
         'format': 'jsonl',
diff --git a/examples/compare/html_extract_compare_v1.py b/examples/compare/html_extract_compare_v1.py
index b69041bd..a45ed890 100644
--- a/examples/compare/html_extract_compare_v1.py
+++ b/examples/compare/html_extract_compare_v1.py
@@ -1,9 +1,14 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/compare/old_new_compare_10000.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/compare/old_new_compare_10000.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/compare/html_extract_compare_v2_example_dataset.py b/examples/compare/html_extract_compare_v2_example_dataset.py
index 6197d466..f4449b2c 100644
--- a/examples/compare/html_extract_compare_v2_example_dataset.py
+++ b/examples/compare/html_extract_compare_v2_example_dataset.py
@@ -27,6 +27,9 @@
 from dingo.config.input_args import InputArgs
 from dingo.exec.base import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 # API 配置
 OPENAI_MODEL = 'deepseek-chat'
 OPENAI_URL = os.getenv("OPENAI_BASE_URL")
@@ -53,7 +56,7 @@ def evaluate_html_extract_compare_dataset():
     # 配置参数
     input_data = {
         "task_name": "html_extract_compare_v2_evaluation",
-        "input_path": str(Path("test/data/html_extract_compare_test.jsonl")),
+        "input_path": str(PROJECT_ROOT / "test/data/html_extract_compare_test.jsonl"),
         "output_path": "output/html_extract_compare_evaluation/",
         # "log_level": "INFO",
 
diff --git a/examples/continue/continue.py b/examples/continue/continue.py
index 5fbe9e8a..889b5495 100644
--- a/examples/continue/continue.py
+++ b/examples/continue/continue.py
@@ -1,10 +1,15 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def exec_first():
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl"
@@ -35,7 +40,7 @@ def exec_first():
 
 def exec_second():
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/core/score.py b/examples/core/score.py
index c886db58..c3502bb7 100644
--- a/examples/core/score.py
+++ b/examples/core/score.py
@@ -5,9 +5,10 @@
 from dingo.model.llm.llm_text_quality_model_base import LLMTextQualityModelBase
 from dingo.model.rule.rule_common import RuleEnterAndSpace
 
-OPENAI_MODEL = 'deepseek-chat'
-OPENAI_URL = 'https://api.deepseek.com/v1'
-OPENAI_KEY = os.getenv("OPENAI_KEY")
+# Configure LLM (set your API key via environment variable OPENAI_KEY)
+OPENAI_MODEL = os.getenv("OPENAI_MODEL", "gpt-4o")
+OPENAI_URL = os.getenv("OPENAI_URL", "https://api.openai.com/v1")
+OPENAI_KEY = os.getenv("OPENAI_KEY", "YOUR_API_KEY")  # Set OPENAI_KEY env var
 
 
 def llm():
diff --git a/examples/custom/sdk_custom_llm.py b/examples/custom/sdk_custom_llm.py
index 8c61810d..ba7ea604 100644
--- a/examples/custom/sdk_custom_llm.py
+++ b/examples/custom/sdk_custom_llm.py
@@ -1,9 +1,22 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
+# Configure LLM (set your API key via environment variable OPENAI_KEY)
+LLM_CONFIG = {
+    "key": os.getenv("OPENAI_KEY", "YOUR_API_KEY"),
+    "api_url": os.getenv("OPENAI_URL", "https://api.openai.com/v1"),
+    "model": os.getenv("OPENAI_MODEL", "gpt-4o")
+}
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
@@ -18,7 +31,7 @@
             {
                 "fields": {"content": "content"},
                 "evals": [
-                    {"name": "LLMTextRepeat", "config": {"key": "", "api_url": ""}},
+                    {"name": "LLMTextRepeat", "config": LLM_CONFIG},
                 ]
             }
         ]
diff --git a/examples/custom/sdk_custom_rule.py b/examples/custom/sdk_custom_rule.py
index e155fb88..79563564 100644
--- a/examples/custom/sdk_custom_rule.py
+++ b/examples/custom/sdk_custom_rule.py
@@ -1,9 +1,14 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_local_json.json",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_json.json"),
         "dataset": {
             "source": "local",
             "format": "json",
diff --git a/examples/dataman/dataman.py b/examples/dataman/dataman.py
index bd51de64..1849815c 100644
--- a/examples/dataman/dataman.py
+++ b/examples/dataman/dataman.py
@@ -1,9 +1,22 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
+# Configure LLM (set your API key via environment variable OPENAI_KEY)
+LLM_CONFIG = {
+    "key": os.getenv("OPENAI_KEY", "YOUR_API_KEY"),
+    "api_url": os.getenv("OPENAI_URL", "https://api.openai.com/v1"),
+    "model": os.getenv("OPENAI_MODEL", "gpt-4o")
+}
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_dataman_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_dataman_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
@@ -20,7 +33,7 @@
             {
                 "fields": {"content": "content"},
                 "evals": [
-                    {"name": "LLMDatamanAssessment", "config": {"key": "", "api_url": ""}},
+                    {"name": "LLMDatamanAssessment", "config": LLM_CONFIG},
                 ]
             }
         ]
diff --git a/examples/document_parser/document_parsing_quality_ocr.py b/examples/document_parser/document_parsing_quality_ocr.py
index 5a61b621..705df288 100644
--- a/examples/document_parser/document_parsing_quality_ocr.py
+++ b/examples/document_parser/document_parsing_quality_ocr.py
@@ -1,9 +1,19 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
+
     input_data = {
-        "input_path": "../../test/data/test_document_OCR_recognize.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_document_OCR_recognize.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
@@ -18,7 +28,7 @@
             {
                 "fields": {"id": "id", "content": "pred_content", "prompt": "gt_markdown"},
                 "evals": [
-                    {"name": "LLMMinerURecognizeQuality", "config": {"key": "", "api_url": ""}},
+                    {"name": "LLMMinerURecognizeQuality", "config": {"key": OPENAI_KEY, "api_url": OPENAI_URL, "model": OPENAI_MODEL}},
                 ]
             }
         ]
diff --git a/examples/document_parser/document_parsing_quality_ocr_train.py b/examples/document_parser/document_parsing_quality_ocr_train.py
index de42b46c..918d9772 100644
--- a/examples/document_parser/document_parsing_quality_ocr_train.py
+++ b/examples/document_parser/document_parsing_quality_ocr_train.py
@@ -1,9 +1,14 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "test/data/test_document_OCR_recognize.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_document_OCR_recognize.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/document_parser/vlm_document_parser_quality.py b/examples/document_parser/vlm_document_parser_quality.py
index ac8b60ef..83c96f2b 100644
--- a/examples/document_parser/vlm_document_parser_quality.py
+++ b/examples/document_parser/vlm_document_parser_quality.py
@@ -1,9 +1,14 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_img_md.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_img_md.jsonl"),
         "dataset": {
             "source": "local",
             "format": "image",
diff --git a/examples/document_parser/vlm_layout_quality.py b/examples/document_parser/vlm_layout_quality.py
index b165223f..d1219919 100644
--- a/examples/document_parser/vlm_layout_quality.py
+++ b/examples/document_parser/vlm_layout_quality.py
@@ -1,9 +1,14 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_layout_quality.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_layout_quality.jsonl"),
         "dataset": {
             "source": "local",
             "format": "image",
diff --git a/examples/factcheck/dataset_factcheck_evaluation.py b/examples/factcheck/dataset_factcheck_evaluation.py
index 479d83fa..dd3d72d3 100644
--- a/examples/factcheck/dataset_factcheck_evaluation.py
+++ b/examples/factcheck/dataset_factcheck_evaluation.py
@@ -16,9 +16,9 @@
 # Force import factuality evaluation modules
 from dingo.model.llm.llm_factcheck_public import LLMFactCheckPublic
 
-OPENAI_MODEL = 'deepseek-chat'
-OPENAI_URL = 'https://api.deepseek.com/v1'
-OPENAI_KEY = os.getenv("OPENAI_KEY")
+OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
 
 
 def evaluate_factuality_jsonl_dataset():
@@ -92,9 +92,8 @@ def evaluate_single_data_example():
     result = evaluator.eval(test_data)
 
     print("\n=== Evaluation Result ===")
-    print(f"Error Status: {result.eval_status}")
-    print(f"Type: {result.type}")
-    print(f"Name: {result.name}")
+    print(f"Error Status: {result.status}")
+    print(f"Label: {result.label}")
     print(f"Reason: {result.reason}")
 
 
diff --git a/examples/hallucination/sdk_hallucination_detection.py b/examples/hallucination/sdk_hallucination_detection.py
index d74119b2..40f292c5 100644
--- a/examples/hallucination/sdk_hallucination_detection.py
+++ b/examples/hallucination/sdk_hallucination_detection.py
@@ -8,15 +8,17 @@
 against provided reference contexts.
 """
 
+import os
+
 from dingo.config.input_args import EvaluatorLLMArgs
 from dingo.io.input import Data
 from dingo.model.llm.llm_hallucination import LLMHallucination
 
-# Configure LLM
+# Configure LLM (set your API key via environment variable OPENAI_KEY)
 LLMHallucination.dynamic_config = EvaluatorLLMArgs(
-    key='sk-xxx',
-    api_url='https://api.deepseek.com',
-    model='deepseek-chat',
+    key=os.getenv("OPENAI_KEY", "YOUR_API_KEY"),  # Replace with your API key or set OPENAI_KEY env var
+    api_url=os.getenv("OPENAI_URL", "https://api.openai.com/v1"),
+    model=os.getenv("OPENAI_MODEL", "gpt-4o"),
 )
 
 
@@ -34,10 +36,8 @@ def example_1_basic_hallucination_detection():
 
     result = LLMHallucination.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"Reason: {result.reason[0]}")
     print(f"Score: {getattr(result, 'score', 'N/A')}")
     print()
@@ -57,10 +57,8 @@ def example_2_no_hallucination():
 
     result = LLMHallucination.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"Reason: {result.reason[0]}")
     print(f"Score: {getattr(result, 'score', 'N/A')}")
     print()
@@ -86,14 +84,9 @@ def example_3_multiple_contexts():
 
     result = LLMHallucination.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"Score: {getattr(result, 'score', 'N/A')}")
-    # print(f"Verdict Details:")
-    # for detail in getattr(result, 'verdict_details', []):
-    #     print(f"  - {detail}")
     print()
 
 
@@ -118,10 +111,8 @@ def example_4_rag_scenario():
 
     result = LLMHallucination.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"Score: {getattr(result, 'score', 'N/A')}")
     print("Detailed Analysis:")
     print(result.reason[0])
@@ -141,16 +132,14 @@ def example_5_missing_context():
 
     result = LLMHallucination.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"Reason: {result.reason[0]}")
     print()
 
 
 def example_6_clear_hallucination():
-    """Example 6: Clear hallucination case that triggers eval_status=True"""
+    """Example 6: Clear hallucination case that triggers status=True"""
     print("=== Example 6: Clear Hallucination (Error Triggered) ===")
 
     # Create a case where the response clearly contradicts multiple contexts
@@ -170,17 +159,10 @@ def example_6_clear_hallucination():
 
     result = LLMHallucination.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
-    print(f"Score: {getattr(result, 'score', 'N/A')}")
-    print("Detailed Analysis:")
-    print(result.reason[0])
-    # if hasattr(result, 'verdict_details'):
-    #     print("Verdict Details:")
-    #     for detail in result.verdict_details:
-    #         print(f"  - {detail}")
+    print(f"Error Status: {result.status}")
+    print(f"Label: {result.label}")
+    print(f"Detailed Analysis:")
+    print(result.reason[0] if result.reason else "N/A")
     print()
 
 
diff --git a/examples/hallucination/sdk_rule_hhem_detection.py b/examples/hallucination/sdk_rule_hhem_detection.py
index 5302a880..576fbdc6 100644
--- a/examples/hallucination/sdk_rule_hhem_detection.py
+++ b/examples/hallucination/sdk_rule_hhem_detection.py
@@ -33,14 +33,12 @@ def example_1_basic_rule_hhem_detection():
 
     result = RuleHallucinationHHEM.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Error Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"HHEM Score: {getattr(result, 'score', 'N/A'):.3f}")
     print(f"Threshold: {RuleHallucinationHHEM.dynamic_config.threshold}")
     print("\nDetailed Analysis:")
-    print(result.reason[0])
+    print(result.reason[0] if result.reason else "N/A")
     print()
 
 
@@ -61,13 +59,11 @@ def example_2_no_hallucination_rule():
 
     result = RuleHallucinationHHEM.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Error Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"HHEM Score: {getattr(result, 'score', 'N/A'):.3f}")
     print("\nDetailed Analysis:")
-    print(result.reason[0])
+    print(result.reason[0] if result.reason else "N/A")
     print()
 
 
@@ -91,13 +87,11 @@ def example_3_complex_scenario_rule():
 
     result = RuleHallucinationHHEM.eval(data)
 
-    print(f"Error Status: {result.eval_status}")
-    # print(f"Type: {result.type}")
-    # print(f"Name: {result.name}")
-    print(f"Type: {result.eval_details}")
+    print(f"Error Status: {result.status}")  # True = hallucination detected, False = no hallucination
+    print(f"Label: {result.label}")
     print(f"HHEM Score: {getattr(result, 'score', 'N/A'):.3f}")
     print("\nDetailed Analysis:")
-    print(result.reason[0])
+    print(result.reason[0] if result.reason else "N/A")
     print()
 
 
@@ -156,7 +150,7 @@ def example_5_batch_evaluation_rule():
 
     print("Batch Rule-based Evaluation Results:")
     for i, result in enumerate(results):
-        print(f"  Item {i + 1}: Error={result.eval_status}, Score={getattr(result, 'score', 'N/A'):.3f}")
+        print(f"  Item {i + 1}: Error={result.status}, Score={getattr(result, 'score', 'N/A'):.3f}")
     print()
 
 
@@ -182,7 +176,7 @@ def example_6_threshold_comparison_rule():
         RuleHallucinationHHEM.dynamic_config.threshold = threshold
         result = RuleHallucinationHHEM.eval(data)
 
-        print(f"Threshold {threshold}: Error={result.eval_status}, Score={getattr(result, 'score', 'N/A'):.3f}")
+        print(f"Threshold {threshold}: Error={result.status}, Score={getattr(result, 'score', 'N/A'):.3f}")
 
     # Restore original threshold
     RuleHallucinationHHEM.dynamic_config.threshold = original_threshold
@@ -211,7 +205,7 @@ def example_7_performance_benchmark_rule():
     end_time = time.time()
 
     print(f"Rule-based HHEM Inference Time: {end_time - start_time:.3f} seconds")
-    print(f"Result: Error={result.eval_status}, Score={getattr(result, 'score', 'N/A'):.3f}")
+    print(f"Result: Error={result.status}, Score={getattr(result, 'score', 'N/A'):.3f}")
     print(f"Model Info: Local HHEM-2.1-Open (Rule-based)")
     print()
 
diff --git a/examples/image/sdk_image.py b/examples/image/sdk_image.py
index 483e0097..d01f94d8 100644
--- a/examples/image/sdk_image.py
+++ b/examples/image/sdk_image.py
@@ -1,13 +1,22 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def image():
     input_data = {
-        "input_path": "../../test/data/test_local_img.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_img.jsonl"),
         "dataset": {
             "source": "local",
             "format": "image",
+            "field": {
+                "id": "id",
+                "image": "img"
+            }
         },
         "executor": {
             "result_save": {
diff --git a/examples/image/sdk_image_label_overlap.py b/examples/image/sdk_image_label_overlap.py
index 39f63265..064156f9 100644
--- a/examples/image/sdk_image_label_overlap.py
+++ b/examples/image/sdk_image_label_overlap.py
@@ -1,10 +1,15 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def image_label_overlap():
     input_data = {
-        "input_path": "../../test/data/img_label/test_img_label_overlap.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/img_label/test_img_label_overlap.jsonl"),
         "dataset": {
             "source": "local",
             "format": "image",
diff --git a/examples/image/sdk_image_label_visualization.py b/examples/image/sdk_image_label_visualization.py
index 98753d59..4a11d153 100644
--- a/examples/image/sdk_image_label_visualization.py
+++ b/examples/image/sdk_image_label_visualization.py
@@ -1,10 +1,15 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def image_label_overlap():
     input_data = {
-        "input_path": "../../test/data/img_label/test_img_label_visualization.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/img_label/test_img_label_visualization.jsonl"),
         "dataset": {
             "source": "local",
             "format": "image",
diff --git a/examples/image/sdk_image_relevant.py b/examples/image/sdk_image_relevant.py
index 11e95a0c..67e2e35a 100644
--- a/examples/image/sdk_image_relevant.py
+++ b/examples/image/sdk_image_relevant.py
@@ -1,10 +1,15 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def image_relevant():
     input_data = {
-        "input_path": "../../test/data/test_img_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_img_jsonl.jsonl"),
         "output_path": "output/hallucination_evaluation/",
         "dataset": {
             "source": "local",
diff --git a/examples/image/sdk_image_repeat.py b/examples/image/sdk_image_repeat.py
index 24bd0cd5..080a5a2c 100644
--- a/examples/image/sdk_image_repeat.py
+++ b/examples/image/sdk_image_repeat.py
@@ -1,10 +1,15 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def image_repeat():
     input_data = {
-        "input_path": "../../test/data/test_img_repeat.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_img_repeat.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/image/sdk_image_text_similar.py b/examples/image/sdk_image_text_similar.py
index 6fcf0ae4..15b0ee45 100644
--- a/examples/image/sdk_image_text_similar.py
+++ b/examples/image/sdk_image_text_similar.py
@@ -1,10 +1,15 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 
 def image_text_similar():
     input_data = {
-        "input_path": "../../test/data/test_img_text.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_img_text.jsonl"),
         "dataset": {
             "source": "local",
             "format": "image",
diff --git a/examples/llm_and_rule/llm_and_rule_mix.py b/examples/llm_and_rule/llm_and_rule_mix.py
index ea35f088..baa21809 100644
--- a/examples/llm_and_rule/llm_and_rule_mix.py
+++ b/examples/llm_and_rule/llm_and_rule_mix.py
@@ -1,15 +1,19 @@
 import os
+from pathlib import Path
 
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
-    OPENAI_MODEL = 'deepseek-chat'
-    OPENAI_URL = 'https://api.deepseek.com/v1'
-    OPENAI_KEY = os.getenv("OPENAI_KEY")
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
 
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/llm_and_rule/llm_local.py b/examples/llm_and_rule/llm_local.py
index 76adefe0..29666e18 100644
--- a/examples/llm_and_rule/llm_local.py
+++ b/examples/llm_and_rule/llm_local.py
@@ -4,6 +4,9 @@
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
 OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
 OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
@@ -16,7 +19,7 @@
 
 if __name__ == '__main__':
     input_data = {
-        "input_path": str(Path("test/data/test_local_jsonl.jsonl")),
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/llm_and_rule/llm_remote.py b/examples/llm_and_rule/llm_remote.py
index d05c43db..5a4eb78b 100644
--- a/examples/llm_and_rule/llm_remote.py
+++ b/examples/llm_and_rule/llm_remote.py
@@ -1,9 +1,19 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
+
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
@@ -18,7 +28,7 @@
             {
                 "fields": {"content": "content"},
                 "evals": [
-                    {"name": "LLMTextRepeat", "config": {"model": "deepseek-chat", "key": "", "api_url": "https://api.deepseek.com/v1"}}
+                    {"name": "LLMTextRepeat", "config": {"model": OPENAI_MODEL, "key": OPENAI_KEY, "api_url": OPENAI_URL}}
                 ]
             }
         ]
diff --git a/examples/llm_and_rule/only_llm.py b/examples/llm_and_rule/only_llm.py
index 1cb17c3c..4d160225 100644
--- a/examples/llm_and_rule/only_llm.py
+++ b/examples/llm_and_rule/only_llm.py
@@ -1,15 +1,19 @@
 import os
+from pathlib import Path
 
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
-    OPENAI_MODEL = 'deepseek-chat'
-    OPENAI_URL = 'http://10.140.54.48:29990/v1'
-    OPENAI_KEY = "EMPTY"
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
 
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/llm_and_rule/only_rule.py b/examples/llm_and_rule/only_rule.py
index 5d6e62aa..9c08b612 100644
--- a/examples/llm_and_rule/only_rule.py
+++ b/examples/llm_and_rule/only_rule.py
@@ -1,11 +1,14 @@
-import os
+from pathlib import Path
 
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/long_video/llm_generate_qa.py b/examples/long_video/llm_generate_qa.py
index 2297df1c..207e3864 100644
--- a/examples/long_video/llm_generate_qa.py
+++ b/examples/long_video/llm_generate_qa.py
@@ -1,9 +1,19 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
+
     input_data = {
-        "input_path": "../../test/data/test_long_video_qa.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_long_video_qa.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
@@ -18,7 +28,7 @@
             {
                 "fields": {"id": "video_id", "content": "summary"},
                 "evals": [
-                    {"name": "LLMLongVideoQa", "config": {"key": "", "api_url": ""}}
+                    {"name": "LLMLongVideoQa", "config": {"key": OPENAI_KEY, "api_url": OPENAI_URL, "model": OPENAI_MODEL}}
                 ]
             }
         ]
diff --git a/examples/meta_rater/sdk_meta_rater_evaluation.py b/examples/meta_rater/sdk_meta_rater_evaluation.py
index 4434014e..7597b156 100644
--- a/examples/meta_rater/sdk_meta_rater_evaluation.py
+++ b/examples/meta_rater/sdk_meta_rater_evaluation.py
@@ -1,9 +1,19 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
+
     input_data = {
-        "input_path": "../../test/data/test_meta_rater.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_meta_rater.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
@@ -18,10 +28,10 @@
             {
                 "fields": {"content": "content"},
                 "evals": [
-                    {"name": "LLMMetaRaterEvaluation", "config": {"key": "", "api_url": ""}},
-                    {"name": "PromptMetaRaterReadability", "config": {"key": "", "api_url": ""}},
-                    {"name": "PromptMetaRaterReasoning", "config": {"key": "", "api_url": ""}},
-                    {"name": "PromptMetaRaterCleanliness", "config": {"key": "", "api_url": ""}},
+                    {"name": "LLMMetaRaterEvaluation", "config": {"key": OPENAI_KEY, "api_url": OPENAI_URL, "model": OPENAI_MODEL}},
+                    {"name": "PromptMetaRaterReadability", "config": {"key": OPENAI_KEY, "api_url": OPENAI_URL, "model": OPENAI_MODEL}},
+                    {"name": "PromptMetaRaterReasoning", "config": {"key": OPENAI_KEY, "api_url": OPENAI_URL, "model": OPENAI_MODEL}},
+                    {"name": "PromptMetaRaterCleanliness", "config": {"key": OPENAI_KEY, "api_url": OPENAI_URL, "model": OPENAI_MODEL}},
                 ]
             }
         ]
diff --git a/examples/multi_turn_dialogues/sdk_mtbench101_llm.py b/examples/multi_turn_dialogues/sdk_mtbench101_llm.py
index 9aeed06c..123c32b1 100644
--- a/examples/multi_turn_dialogues/sdk_mtbench101_llm.py
+++ b/examples/multi_turn_dialogues/sdk_mtbench101_llm.py
@@ -1,12 +1,16 @@
 import os
+from pathlib import Path
 
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
-    OPENAI_MODEL = 'deepseek-chat'
-    OPENAI_URL = 'https://api.deepseek.com/v1'
-    OPENAI_KEY = os.getenv("OPENAI_KEY")
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
     common_config = {
         "model": OPENAI_MODEL,
         "key": OPENAI_KEY,
@@ -14,7 +18,7 @@
     }
 
     input_data = {
-        "input_path": "../../test/data/test_mtbench101_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_mtbench101_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "multi_turn_dialog",
diff --git a/examples/multi_turn_dialogues/sdk_mtbench101_rule_all.py b/examples/multi_turn_dialogues/sdk_mtbench101_rule_all.py
index 6b6327a3..fa7c4d00 100644
--- a/examples/multi_turn_dialogues/sdk_mtbench101_rule_all.py
+++ b/examples/multi_turn_dialogues/sdk_mtbench101_rule_all.py
@@ -1,9 +1,14 @@
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
     input_data = {
-        "input_path": "../../test/data/test_mtbench101_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_mtbench101_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "multi_turn_dialog",
diff --git a/examples/rag/dataset_rag_eval_baseline.py b/examples/rag/dataset_rag_eval_baseline.py
index bd1cc791..7eac9731 100644
--- a/examples/rag/dataset_rag_eval_baseline.py
+++ b/examples/rag/dataset_rag_eval_baseline.py
@@ -32,6 +32,9 @@
 from dingo.exec import Executor
 from dingo.io.output.summary_model import SummaryModel
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 # 配置（从环境变量读取）
 OPENAI_MODEL = os.getenv("OPENAI_MODEL", "gpt-4o-mini")
 OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.openai.com/v1")
@@ -39,7 +42,7 @@
 EMBEDDING_MODEL = os.getenv("EMBEDDING_MODEL", "text-embedding-3-large")
 
 # 数据文件路径
-INPUT_DATA_PATH = str(Path("test/data/fiqa.jsonl"))  # 或 "test/data/ragflow_eval_data_50.jsonl"
+INPUT_DATA_PATH = str(PROJECT_ROOT / "test/data/fiqa.jsonl")  # 或 "test/data/ragflow_eval_data_50.jsonl"
 
 
 def print_metrics_summary(summary: SummaryModel):
diff --git a/examples/rag/sdk_rag_eval.py b/examples/rag/sdk_rag_eval.py
index c6ef154f..64c1487f 100644
--- a/examples/rag/sdk_rag_eval.py
+++ b/examples/rag/sdk_rag_eval.py
@@ -17,10 +17,10 @@
 from dingo.model.llm.rag.llm_rag_context_relevancy import LLMRAGContextRelevancy
 from dingo.model.llm.rag.llm_rag_faithfulness import LLMRAGFaithfulness
 
-# 配置（从环境变量读取，或直接设置）
+# 配置（从环境变量读取）
 OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
 OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com")
-OPENAI_KEY = os.getenv("OPENAI_API_KEY", "YOUR_API_KEY")
+OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
 
 
 def test_faithfulness():
diff --git a/examples/register/sdk_register_llm.py b/examples/register/sdk_register_llm.py
index c28ea179..647c77be 100644
--- a/examples/register/sdk_register_llm.py
+++ b/examples/register/sdk_register_llm.py
@@ -1,11 +1,15 @@
 import os
+from pathlib import Path
 
 from dingo.model import Model
 from dingo.model.llm.base_openai import BaseOpenAI
 
-OPENAI_MODEL = 'deepseek-chat'
-OPENAI_URL = 'https://api.deepseek.com/v1'
-OPENAI_KEY = os.getenv("OPENAI_KEY")
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
+OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
 
 common_config = {
     "model": OPENAI_MODEL,
@@ -30,7 +34,7 @@ class LlmTextQualityRegister(BaseOpenAI):
     from dingo.exec import Executor
 
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
diff --git a/examples/register/sdk_register_rule.py b/examples/register/sdk_register_rule.py
index 4b33f3de..30626a32 100644
--- a/examples/register/sdk_register_rule.py
+++ b/examples/register/sdk_register_rule.py
@@ -24,11 +24,16 @@ def eval(cls, input_data: Data) -> EvalDetail:
 
 
 if __name__ == '__main__':
+    from pathlib import Path
+
     from dingo.config import InputArgs
     from dingo.exec import Executor
 
+    # 获取项目根目录
+    PROJECT_ROOT = Path(__file__).parent.parent.parent
+
     input_data = {
-        "input_path": "../../test/data/test_local_json.json",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_json.json"),
         "dataset": {
             "source": "local",
             "format": "json",
diff --git a/examples/security/text_security_politics.py b/examples/security/text_security_politics.py
index 0d0cd9bd..2737a1c3 100644
--- a/examples/security/text_security_politics.py
+++ b/examples/security/text_security_politics.py
@@ -1,9 +1,19 @@
+import os
+from pathlib import Path
+
 from dingo.config import InputArgs
 from dingo.exec import Executor
 
+# 获取项目根目录
+PROJECT_ROOT = Path(__file__).parent.parent.parent
+
 if __name__ == '__main__':
+    OPENAI_MODEL = os.getenv("OPENAI_MODEL", "deepseek-chat")
+    OPENAI_URL = os.getenv("OPENAI_BASE_URL", "https://api.deepseek.com/v1")
+    OPENAI_KEY = os.getenv("OPENAI_API_KEY", "")
+
     input_data = {
-        "input_path": "../../test/data/test_local_jsonl.jsonl",
+        "input_path": str(PROJECT_ROOT / "test/data/test_local_jsonl.jsonl"),
         "dataset": {
             "source": "local",
             "format": "jsonl",
@@ -18,7 +28,7 @@
             {
                 "fields": {"content": "content"},
                 "evals": [
-                    {"name": "LLMSecurityPolitics", "config": {"key": "", "api_url": ""}}
+                    {"name": "LLMSecurityPolitics", "config": {"key": OPENAI_KEY, "api_url": OPENAI_URL, "model": OPENAI_MODEL}}
                 ],
             }
         ]