fix: cross-platform path handling and config group isolation

Author: mmmay0722

README.md

@@ -105,17 +105,21 @@ uv run webqa-agent init -m gen
 # Edit config.yaml: target.url, llm_config.api_key
 # Configure test_config
 # For more details, see "Usage > Generate Mode - Configuration" below
-uv run webqa-agent gen      # Run Generate Mode
-uv run webqa-agent gen -c /path/to/config.yaml -w 4      # Generate Mode with specified config and 4 parallel workers
+# Run Generate Mode
+uv run webqa-agent gen
+# Generate Mode with specified config and 4 parallel workers
+uv run webqa-agent gen -c /path/to/config.yaml -w 4
 
 # 4) Run Mode
 # Initialize Run mode configuration (config_run.yaml)
 uv run webqa-agent init -m run
 # Edit config.yaml: target.url, llm_config.api_key
 # Write natural language test cases
 # For more details, see "Usage > Run Mode - Configuration" below
-uv run webqa-agent run     # Run Run Mode
-uv run webqa-agent run -c /path/to/config_run.yaml -w 4     # Run Mode with specified config and 4 parallel workers
+# Run Run Mode
+uv run webqa-agent run
+# Run Mode with specified config and 4 parallel workers
+uv run webqa-agent run -c /path/to/config_run.yaml -w 4
 ```
 
 ### 🔧 Generate Mode - Optional Dependencies
@@ -166,7 +170,6 @@ For more details, please refer to [docs/MODES&CLI.md](docs/MODES&CLI.md)
 target:
   url: https://example.com              # Website URL to test
   description: Website QA testing
-  max_concurrent_tests: 2               # Optional, default 2
 
 test_config:
   function_test:                        # Functional testing
@@ -185,20 +188,13 @@ llm_config:                             # LLM configuration, supports OpenAI, An
   filter_model: gpt-4o-mini             # Lightweight model for element filtering (optional)
   api_key: your_api_key                 # Or set via environment variable (OPENAI_API_KEY)
   base_url: https://api.openai.com/v1   # Optional, API endpoint. For OpenAI-compatible models (Doubao, Qwen, etc.), set to their API endpoint
-  temperature: 0.1                      # Optional, model temperature
-  # For detailed configuration examples (OpenAI, Claude, Gemini) and reasoning settings,
-  # see config/config.yaml.example
 
 browser_config:
-  viewport: {"width": 1280, "height": 720}
   headless: False                       # Auto True in Docker
   language: en-US
 
 report:
   language: en-US                       # zh-CN or en-US
-
-log:
-  level: info                           # debug, info, warning, error
 ```
 
 ### Run Mode - Configuration
@@ -214,7 +210,6 @@ For more details and test case writing specifications, please refer to [docs/MOD
 ```yaml
 target:
   url: https://example.com              # Target website URL
-  max_concurrent_tests: 2               # Maximum concurrent test count
 
 llm_config:                             # LLM configuration
   api: openai

README_zh-CN.md

@@ -105,17 +105,21 @@ uv run webqa-agent init -m gen
 # 编辑 config.yaml：target.url、llm_config.api_key
 # 配置 test_config
 # 更多说明见下方“使用说明 > Gen 模式 - 配置介绍”
-uv run webqa-agent gen      # 运行 Gen 模式
-uv run webqa-agent gen -c /path/to/config.yaml -w 4      # 指定配置文件路径，使用 4 个并行 Worker
+# 运行 Gen 模式
+uv run webqa-agent gen
+# 指定配置文件路径，使用 4 个并行 Worker
+uv run webqa-agent gen -c /path/to/config.yaml -w 4
 
 # 4) Run 模式
 # 初始化 Run 模式配置 (config_run.yaml)
 uv run webqa-agent init -m run
 # 编辑 config.yaml：target.url、llm_config.api_key
 # 编写自然语言用例
 # 更多说明见下方“使用说明 > Run 模式 - 配置介绍”
-uv run webqa-agent run     # 运行 Run 模式
-uv run webqa-agent run -c /path/to/config.yaml -w 4      # 指定配置文件路径，使用 4 个并行 Worker
+# 运行 Run 模式
+uv run webqa-agent run
+# 指定配置文件路径，使用 4 个并行 Worker
+uv run webqa-agent run -c /path/to/config.yaml -w 4
 ```
 
 ### 🔧 Generate 模式 - 可选依赖
@@ -166,7 +170,6 @@ curl -fsSL https://raw.githubusercontent.com/MigoXLab/webqa-agent/main/start.sh
 target:
   url: https://example.com              # 需要测试的网站 URL
   description: 网站质量保证测试
-  max_concurrent_tests: 2               # 可选，默认并发数 2
 
 test_config:
   function_test:                        # 功能测试
@@ -185,22 +188,13 @@ llm_config:                             # LLM 配置，支持 OpenAI、Anthropic
   filter_model: gpt-4o-mini             # 轻量级模型用于元素过滤（可选）
   api_key: your_api_key                 # 或通过环境变量设置 (OPENAI_API_KEY)
   base_url: https://api.openai.com/v1   # 可选，API 端点。对于 OpenAI 兼容格式模型（豆包、通义千问等），设置为对应的 API 端点
-  temperature: 0.1                      # 可选，
-  # 详细的配置示例（OpenAI、Claude、Gemini）和推理设置说明，
-  # 请参考 config/config.yaml.example
 
 browser_config:
-  viewport: {"width": 1280, "height": 720}
   headless: False                       # Docker 环境自动设为 True
   language: en-US
-  cookies: []
-  save_screenshots: False
 
 report:
   language: en-US                       # zh-CN 或 en-US
-
-log:
-  level: info                           # debug, info, warning, error
 ```
 
 ### Run 模式 - 配置介绍
@@ -216,7 +210,6 @@ Run 模式配置文件需包含 `cases` 字段，用于定义具体的测试用
 ```yaml
 target:
   url: https://example.com              # 目标网站 URL
-  max_concurrent_tests: 2               # 最大并发测试数
 
 llm_config:                             # LLM 配置
   api: openai

webqa_agent/actions/action_handler.py

@@ -8,7 +8,7 @@
 import re
 from contextvars import ContextVar
 from dataclasses import dataclass, field
-from pathlib import Path
+from pathlib import Path, PurePosixPath
 from typing import Any, Dict, List, Optional, Union
 
 from playwright.async_api import Frame, Page
@@ -1733,9 +1733,8 @@ async def b64_page_screenshot(
                 file_path = session_dir / filename
                 file_path_str = str(file_path)
 
-                # Return path relative to the report root for HTML rendering
-                # Screenshots are stored in report_dir/screenshots/ and report is in report_dir/run_report.html
-                relative_path = os.path.join(session_dir.name, filename)
+                # Use POSIX separators to keep HTML assets valid on Windows/macOS/Linux.
+                relative_path = str(PurePosixPath(session_dir.name) / filename)
 
             # Capture screenshot (always returns bytes)
             screenshot_bytes = await self.take_screenshot(

webqa_agent/executor/case_mode.py

@@ -242,15 +242,33 @@ async def run(
                 logging.error(f"{icon['cross']} {friendly_msg}")
                 raise ValueError(friendly_msg) from e
 
-            # Initialize case executor with explicit report_dir
-            case_executor = CaseExecutor(
-                test_config=test_config,
-                llm_config=llm_config,
-                report_dir=report_dir
-            )
+            case_results = []
+            case_executor = None
 
-            # Execute all cases
-            case_results = await case_executor.execute_cases(cases=cases, workers=workers)
+            # Split cases by originating config in order to isolate browser sessions between groups
+            case_groups = []
+            start_idx = 0
+            for config in configs:
+                cfg_cases = config.get('cases', [])
+                if not cfg_cases:
+                    continue
+                end_idx = start_idx + len(cfg_cases)
+                case_groups.append(cases[start_idx:end_idx])
+                start_idx = end_idx
+
+            for group_index, group_cases in enumerate(case_groups, start=1):
+                logging.info(f'Executing config #{group_index} ({len(group_cases)} case(s)) with isolated sessions')
+                case_executor = CaseExecutor(
+                    test_config=test_config,
+                    llm_config=llm_config,
+                    report_dir=report_dir,
+                )
+
+                group_results = await case_executor.execute_cases(cases=group_cases, workers=workers)
+                if group_results:
+                    case_results.extend(group_results)
+                else:
+                    logging.warning(f'No results returned for config #{group_index}')
 
             if case_results is None:
                 logging.warning('case_executor.execute_cases returned None, treating as empty list')

webqa_agent/executor/result_aggregator.py

@@ -203,7 +203,13 @@ def _add_files_from_dir(directory: Path):
                         index_data['count'] = count
                         aggregated[mode]['index'] = index_data
         except Exception as backfill_err:
-            logging.warning(f'Failed to backfill aggregated results: {backfill_err}')
+                logging.warning(f'Failed to backfill aggregated results: {backfill_err}')
+
+        # Normalize any file-system paths to POSIX style so HTML works cross-platform
+        try:
+            aggregated = self._normalize_paths_for_web(aggregated)
+        except Exception as norm_err:
+            logging.warning(f'Failed to normalize paths for web: {norm_err}')
 
         # Write the aggregated results to test_results.json
         output_path = os.path.join(report_dir, 'test_results.json')
@@ -293,6 +299,29 @@ def _serialize_data_for_inline(data: Any) -> str:
         )
         return safe
 
+    @staticmethod
+    def _normalize_paths_for_web(data: Any) -> Any:
+        """Convert screenshot paths to web-friendly separators.
+
+        Windows paths use backslashes, which break relative asset loading in
+        the generated HTML. This recursively walks the data structure and
+        rewrites strings that look like screenshot paths to POSIX style.
+        """
+
+        def _normalize(value: Any) -> Any:
+            if isinstance(value, dict):
+                normalized = {k: _normalize(v) for k, v in value.items()}
+                if normalized.get('type') == 'path' and isinstance(normalized.get('data'), str):
+                    normalized['data'] = normalized['data'].replace('\\', '/')
+                return normalized
+            if isinstance(value, list):
+                return [_normalize(item) for item in value]
+            if isinstance(value, str) and 'screenshots' in value:
+                return value.replace('\\', '/')
+            return value
+
+        return _normalize(data)
+
     def generate_html_report_fully_inlined(self, test_session, report_dir: str | None = None, aggregated_data: Dict[str, Any] = None) -> str:
         """Generate a fully inlined HTML report for the test session."""
         import json
@@ -329,6 +358,10 @@ def generate_html_report_fully_inlined(self, test_session, report_dir: str | Non
 
             if data is None:
                 data = test_session.to_dict()
+
+            # Ensure any Windows-style paths are safe for browser consumption
+            data = self._normalize_paths_for_web(data)
+
             safe_data_json = self._serialize_data_for_inline(data)
             datajs_content = f'window.testResultData = {safe_data_json};'