fix: Codex 全量审查修正

Critical:
- C1: assemble_mode 'plain' → 'ids'（合法枚举值）
- C2: question_id 加入 task_type slug 避免跨题型冲突

Important/Minor:
- generate_one 移除未用的 embed_fn/similarity_threshold 参数
- config.py 注释 11→12 同步
- 测试 question_id 断言更新

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

tools/generate_questions.py

@@ -594,7 +594,7 @@ async def _run_inference_for_questions(
                 ocr=None,
                 verify_vision=False,
                 anchor=False,
-                assemble_mode="plain",
+                assemble_mode="ids",
             )
             result = await run_inference(
                 group,
@@ -861,7 +861,6 @@ async def _run_generate(args: argparse.Namespace) -> None:
 
     async def _generate_with_sem(
         vlm_client,
-        embed_fn_inner,
         tree,
         video_id,
         task_type,
@@ -870,7 +869,6 @@ async def _run_generate(args: argparse.Namespace) -> None:
         exemplars,
         used_node_ids,
         max_retries_inner,
-        similarity_threshold_inner,
         rng_inner,
         session_id,
     ):
@@ -878,7 +876,6 @@ async def _run_generate(args: argparse.Namespace) -> None:
         async with sem:
             return await generate_one(
                 vlm_client,
-                embed_fn_inner,
                 tree,
                 video_id,
                 task_type,
@@ -886,7 +883,6 @@ async def _run_generate(args: argparse.Namespace) -> None:
                 exemplars=exemplars,
                 used_node_ids=used_node_ids,
                 max_retries=max_retries_inner,
-                similarity_threshold=similarity_threshold_inner,
                 rng=rng_inner,
                 session_id=session_id,
             )
@@ -929,7 +925,6 @@ async def _run_generate(args: argparse.Namespace) -> None:
             for _attempt in range(max_retries):
                 candidate = await _generate_with_sem(
                     vlm,
-                    embed_fn,
                     tree,
                     video_id,
                     task_type,
@@ -937,7 +932,6 @@ async def _run_generate(args: argparse.Namespace) -> None:
                     exemplars=exemplars,
                     used_node_ids=used_node_ids,
                     max_retries_inner=1,
-                    similarity_threshold_inner=similarity_threshold,
                     rng_inner=rng,
                     session_id=session_id,
                 )