feat(eval): memoir A/B chapter judging and eval-web parity with dialogue

- Judge baseline excerpt and library chapter separately; build_memoir_compare_summary for gate, nine-dim and leaf deltas.

- Memoir SSE chapter payload: baseline_judge, compare_summary, baseline_judge_error.

- MemoirJudgeOutput: loose score coercion and post-validate clamp; memoir judge prompt caps from settings.

- app-eval-web: two-column MemoirScoreCard layout, MemoirCompareSummary, chapter blocks and CSS.

- Add memoir_compare_summary, log_events, celery_log_context, memoir_pipeline_progress; tests and migration 0014.

- Misc: memory/evidence and enrichment paths, task/orchestrator updates, internal-eval docs, env examples.

api/app/features/evaluation/schemas.py

@@ -134,6 +134,19 @@ class MemoirSubmitOut(BaseModel):
     elapsed_ms: int | None = Field(default=None, ge=0)
 
 
+class MemoirPipelineRunOut(BaseModel):
+    """Redis 流水线快照（memoir_pipeline_run:*）；字段随迭代扩展。"""
+
+    model_config = ConfigDict(extra="allow")
+
+    memoir_correlation_id: str
+    user_id: str | None = None
+    started_at_utc: str | None = None
+    phase1: dict[str, Any] | None = None
+    phase2: list[Any] = Field(default_factory=list)
+    fanout: dict[str, Any] = Field(default_factory=dict)
+
+
 class ManualJudgeConversationBody(BaseModel):
     conversation_id: str
     """与当前评测台选中的 MD 一致，供基准 transcript / 整体打分。"""
@@ -202,8 +215,15 @@ class ManualJudgeMemoirBody(BaseModel):
 
 class ManualJudgeMemoirOut(BaseModel):
     user_id: str
+    judge_provider: EvalJudgeProviderLiteral = "zhipu"
+    judge_model: str = ""
+    """本次请求实际解析后的模型 id（与 `build_eval_judge_llm_spec` 一致）。"""
     chapter_results: list[dict[str, Any]] = Field(default_factory=list)
     story_results: list[dict[str, Any]] = Field(default_factory=list)
+    errors: list[str] = Field(default_factory=list)
+    """单条章节/故事评审或列表加载失败时的可读原因（HTTP 仍为 200）。"""
+    warnings: list[str] = Field(default_factory=list)
+    """无失败但未评到任何条目时的提示（例如成稿均为空）。"""
 
 
 class MemoirChapterSnapOut(BaseModel):