feat: make contexts optional in /api/score

When contexts is absent, metrics that require retrieved_contexts
(faithfulness, context_recall, context_precision, noise_sensitivity)
are automatically skipped and appear in skipped_metrics.
Only answer_relevancy, factual_correctness, semantic_similarity
remain computable without contexts.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

webapp/api/score.py

@@ -73,7 +73,8 @@ def score_sample(
     用于日志记录、质量监控或触发 Agent 自我改进流程。
 
     **contexts 格式**：多个检索片段用 `context_separator`（默认 `" |||| "`）拼接为一个字符串，
-    服务端自动拆分后传入 RAGAS 管道。
+    服务端自动拆分后传入 RAGAS 管道。**contexts 为可选字段**，缺失时自动跳过依赖检索内容的指标
+    （`faithfulness`、`context_recall`、`context_precision`、`noise_sensitivity`）。
 
     **ground_truth 可选**：
     - 提供时：所有指定指标均参与计算。
@@ -99,12 +100,13 @@ def score_sample(
     """
     client = f"{raw_request.client.host}:{raw_request.client.port}" if raw_request.client else "unknown"
     logger.info(
-        "[score] incoming  client=%s  method=%s  content_type=%s  metrics=%s  has_gt=%s",
+        "[score] incoming  client=%s  method=%s  content_type=%s  metrics=%s  has_gt=%s  has_ctx=%s",
         client,
         raw_request.method,
         raw_request.headers.get("content-type", ""),
         request.metrics,
         request.ground_truth is not None,
+        bool(request.contexts),
     )
     settings = _get_settings()