style: format distractor_selector

app/question_gen/distractor_selector.py

@@ -87,9 +87,7 @@ def _select_in_interval(
     upper = correct_score - delta_low
     lower = correct_score - delta_high
     eligible = [
-        (c, s)
+        (c, s) for c, s in zip(candidates, candidate_scores, strict=True) if lower <= s <= upper
-        for c, s in zip(candidates, candidate_scores, strict=True)
-        if lower <= s <= upper
     ]
     if len(eligible) < 3:
         return None
@@ -134,12 +132,20 @@ def _parse_json_object(raw: str) -> dict:
 
 
 async def _generate_pool(
-    vlm: VLMProvider, question: str, correct_text: str,
+    vlm: VLMProvider,
-    material: MaterialContext, n: int, *, session_id: str,
+    question: str,
+    correct_text: str,
+    material: MaterialContext,
+    n: int,
+    *,
+    session_id: str,
 ) -> list[str]:
     """VLM 生成 n 个候选干扰项文本。"""
     system = _load_prompt("ar_distractor_pool.md")
-    user = _material_context_block(question, correct_text, material) + f"\n## N\nGenerate exactly {n} distractors."
+    user = (
+        _material_context_block(question, correct_text, material)
+        + f"\n## N\nGenerate exactly {n} distractors."
+    )
     messages = [{"role": "system", "content": system}, {"role": "user", "content": user}]
     resp = await vlm.chat_with_images(messages, list(material.frame_paths), session_id=session_id)
     # graceful 降级：解析爆炸时返回空池，让退火/hard_fail 接管（区别于契约违反的打分校验）
@@ -164,8 +170,12 @@ async def _generate_pool(
 
 
 async def _score_options(
-    vlm: VLMProvider, question: str, options: list[str],
+    vlm: VLMProvider,
-    material: MaterialContext, *, session_id: str,
+    question: str,
+    options: list[str],
+    material: MaterialContext,
+    *,
+    session_id: str,
 ) -> list[float]:
     """VLM 对 options（首个为正解）逐一打视觉可信度分 [0,1]，返回对齐分数列表。"""
     system = _load_prompt("ar_distractor_score.md")
@@ -211,7 +221,9 @@ async def build_grounded_options(
         vlm, question, correct_text, material, config.candidate_pool_size, session_id=session_id
     )
     # options[0] 恒为正解
-    scored = await _score_options(vlm, question, [correct_text, *candidates], material, session_id=session_id)
+    scored = await _score_options(
+        vlm, question, [correct_text, *candidates], material, session_id=session_id
+    )
     correct_score, cand_scores = scored[0], scored[1:]
     anneal_rounds = 0
 
@@ -250,9 +262,7 @@ async def build_grounded_options(
     hard_fail = chosen is None
     observation: dict[str, object] = {
         "correct_score": correct_score,
-        "chosen": [
+        "chosen": [cand_scores[candidates.index(c)] for c in (chosen or [])],
-            cand_scores[candidates.index(c)] for c in (chosen or [])
-        ],
         "pool_size": len(candidates),
         "anneal_rounds": anneal_rounds,
         "delta_high_final": delta_high,
@@ -261,7 +271,9 @@ async def build_grounded_options(
     if hard_fail:
         logger.warning(
             "grounded selector 硬失败: correct={:.3f}, pool={}, anneal={}",
-            correct_score, len(candidates), anneal_rounds,
+            correct_score,
+            len(candidates),
+            anneal_rounds,
         )
         # observation 仍返回，供 pipeline 落 selector_scores（设计 §3.3 退化观测）
         return SelectorOutcome(observation=observation)