fix: load real tree_data for training-loop diagnosis (runner)

算法保真 §4.7#12（训练循环编排）：仅将 _run_diagnosis 内 tree_data={} 替换为按 batch question_ids 涉及 video 加载真实树注入 run_diagnosis，未改动三级嵌套/ 慢更新/断点续训/gate 等编排逻辑。覆盖 question_ids=None 全量诊断与 question_ids=[...] batch 两条路径。
2026-07-15 22:53:34 -04:00
parent 8dedd082ff
commit 224bb25bb5
2 changed files with 189 additions and 1 deletions
@@ -2174,10 +2174,19 @@ class Runner:
        skill_store = VersionedSkillStore(self._paths.skills_dir)
        diagnose_prompts = self._load_diagnose_prompts()

+        from app.harness.tree_nodes import load_tree_data_for_videos
+
+        if question_ids is not None:
+            qid_set = set(question_ids)
+            video_ids = [q.video_id for q in questions if q.question_id in qid_set]
+        else:
+            video_ids = [q.video_id for q in questions]
+        tree_data = load_tree_data_for_videos(Path(self._config.store_dir), video_ids)
+
        return await run_diagnosis(
            run_id=run_id,
            questions=questions,
-            tree_data={},  # tree_data 由诊断管线内部按需加载
+            tree_data=tree_data,
            llm=self._llm,
            run_log=run_log,
            skill_store=skill_store,