fix(validation): populate screenshot_paths column + backfill from raw_response

Two bugs preventing screenshot display in report UI:

1. Plugin's ValidationRecord constructor did NOT include
   screenshot_paths (plural, JSON array). The column stayed NULL
   even though paths were stored in raw_response. Records created
   by the plugin now include screenshot_paths=webp_paths|jpeg_paths.

2. Backward compat for existing records: _record_to_dict now
   falls back to extracting screenshot_paths from raw_response JSON
   when the DB column is empty.

Also fixes: broken import statement in validation_service.py
(missing closing paren + missing imports from earlier edit).

backend/src/plugins/llm_analysis/plugin.py

@@ -478,6 +478,7 @@ class DashboardValidationPlugin(PluginBase):
                 summary=validation_result.summary,
                 issues=[issue.model_dump() for issue in validation_result.issues],
                 screenshot_path=validation_result.screenshot_path,
+                screenshot_paths=webp_paths or jpeg_paths,
                 raw_response=json.dumps(result_payload, ensure_ascii=False),
             )
             db.add(db_record)

backend/src/services/validation_service.py

@@ -22,6 +22,24 @@ from ..schemas.validation import (
     ValidationTaskResponse,
     ValidationTaskUpdate,
 )
+
+# #region _extract_screenshot_paths [C:2] [TYPE Function]
+# @BRIEF Extract screenshot_paths from raw_response JSON when the column is empty (backward compat).
+def _extract_screenshot_paths(raw_response):
+    if not raw_response:
+        return []
+    import json
+    try:
+        parsed = json.loads(raw_response)
+        paths = parsed.get("screenshot_paths") or []
+        if isinstance(paths, list):
+            return paths
+    except (json.JSONDecodeError, TypeError):
+        pass
+    return []
+
+
+# #endregion _extract_screenshot_paths
 from .llm_provider import LLMProviderService
 
 
@@ -297,7 +315,9 @@ class ValidationTaskService:
             "issues": record.issues or [],
             "raw_response": record.raw_response,
             "screenshot_path": record.screenshot_path,
-            "screenshot_paths": record.screenshot_paths or [],
+            # screenshot_paths column may be empty for runs created before the fix;
+            # fall back to extracting from raw_response JSON
+            "screenshot_paths": record.screenshot_paths or _extract_screenshot_paths(record.raw_response),
             "execution_path": record.execution_path,
             "dataset_health": record.dataset_health,
             "chart_data_results": record.chart_data_results or [],