Delete change_lineage.py

change_lineage.py

@@ -1,151 +0,0 @@
-import json
-import pickle
-import numpy as np
-
-
-def convert_lineage_json_to_genealogy_pkl(
-    lineage_json_path: str,
-    save_pkl_path: str,
-    max_children: int = 4,
-):
-    """
-    将 merge 代码生成的 lineage.json 转换为
-    build_sequences.py 需要的 genealogy.pkl。
-
-    lineage.json 格式（子→父，merge方向）：
-        lineage["L1"][i] = [p0, p1, ...]  # L1[i] 由 L0 中这些点 merge 来
-        lineage["L2"][i] = [p0, p1, ...]  # L2[i] 由 L1 中这些点 merge 来
-        lineage["L3"][i] = [p0, p1, ...]  # L3[i] 由 L2 中这些点 merge 来
-
-    genealogy.pkl 格式（父→子，split方向）：
-        genealogy[1]['children_ids']  shape (N_L0, 4)  # L0每个点在L1中的子节点
-        genealogy[2]['children_ids']  shape (N_L1, 4)  # L1每个点在L2中的子节点
-        genealogy[3]['children_ids']  shape (N_L2, 4)  # L2每个点在L3中的子节点
-
-    注意：merge 是多→一（多个父点合成一个子点），
-          split 是一→多（一个父点分裂出多个子点），
-          所以这里的"子节点"含义是：
-            L0[p] 参与 merge 成了 L1[i]，则 L1[i] 就是 L0[p] 的一个"子节点"
-    """
-    print(f"[convert] 加载 {lineage_json_path}")
-    with open(lineage_json_path, 'r') as f:
-        lineage = json.load(f)
-
-    genealogy = {}
-
-    level_map = {
-        "L1": 1,   # genealogy key=1 对应 L0→L1
-        "L2": 2,   # genealogy key=2 对应 L1→L2
-        "L3": 3,   # genealogy key=3 对应 L2→L3
-    }
-
-    for level_name, gen_key in level_map.items():
-        if level_name not in lineage:
-            print(f"[convert] 警告：lineage 中无 {level_name}，跳过")
-            continue
-
-        merge_list = lineage[level_name]   # list of list，长度 = 该级点数
-        n_children = len(merge_list)       # 子尺度（更粗）点数
-
-        # 父尺度点数：取所有父索引的最大值+1
-        all_parent_indices = [idx for parents in merge_list for idx in parents]
-        if not all_parent_indices:
-            print(f"[convert] 警告：{level_name} 族谱为空，跳过")
-            continue
-        n_parents = max(all_parent_indices) + 1
-
-        print(f"[convert] {level_name}：父尺度点数={n_parents}，子尺度点数={n_children}")
-
-        # 初始化 children_ids，-1 为空位
-        children_ids = np.full((n_parents, max_children), fill_value=-1, dtype=np.int32)
-        child_count  = np.zeros(n_parents, dtype=np.int32)
-        skipped      = 0
-
-        # merge_list[child_idx] = [parent_idx, ...]
-        # 反转：对每个 parent_idx，把 child_idx 填入其 children_ids
-        for child_idx, parent_indices in enumerate(merge_list):
-            for p_idx in parent_indices:
-                if child_count[p_idx] < max_children:
-                    children_ids[p_idx, child_count[p_idx]] = child_idx
-                    child_count[p_idx] += 1
-                else:
-                    skipped += 1
-
-        # 统计
-        has_children     = child_count > 0
-        childless        = (~has_children).sum()
-        avg_child        = child_count[has_children].mean() if has_children.any() else 0
-
-        print(f"           有子节点的父点：{has_children.sum()}/{n_parents}")
-        print(f"           无子节点的父点（孤立点，不会生成序列）：{childless}")
-        print(f"           平均子节点数：{avg_child:.2f}")
-        if skipped:
-            print(f"           ⚠️  因超过 {max_children} 个子节点被截断：{skipped} 条")
-
-        genealogy[gen_key] = {
-            'children_ids': children_ids,   # (N_parents, max_children)
-        }
-
-    # 保存
-    with open(save_pkl_path, 'wb') as f:
-        pickle.dump(genealogy, f, protocol=4)
-
-    size_mb = os.path.getsize(save_pkl_path) / 1024 / 1024
-    print(f"\n[convert] 族谱已保存 → {save_pkl_path}  ({size_mb:.2f} MB)")
-    return genealogy
-
-
-# ── 快速验证 ──────────────────────────────────────────────────
-
-def verify_conversion(lineage_json_path: str, genealogy_pkl_path: str, n_sample: int = 5):
-    """
-    抽查几个点，确认转换逻辑正确：
-    lineage["L1"][child_idx] 里的每个 parent_idx
-    都应该出现在 genealogy[1]['children_ids'][parent_idx] 中。
-    """
-    import os
-    with open(lineage_json_path, 'r') as f:
-        lineage = json.load(f)
-    with open(genealogy_pkl_path, 'rb') as f:
-        genealogy = pickle.load(f)
-
-    print("\n[verify] 抽查转换正确性...")
-    for level_name, gen_key in [("L1", 1), ("L2", 2), ("L3", 3)]:
-        if level_name not in lineage or gen_key not in genealogy:
-            continue
-
-        merge_list   = lineage[level_name]
-        children_ids = genealogy[gen_key]['children_ids']
-        errors       = 0
-
-        sample_indices = np.random.choice(len(merge_list), min(n_sample, len(merge_list)), replace=False)
-        for child_idx in sample_indices:
-            parent_indices = merge_list[child_idx]
-            for p_idx in parent_indices:
-                row = children_ids[p_idx]
-                if child_idx not in row:
-                    print(f"  ❌ {level_name} child={child_idx} 的父节点 p={p_idx} "
-                          f"的 children_ids={row}，未包含 {child_idx}")
-                    errors += 1
-
-        if errors == 0:
-            print(f"  ✅ {level_name}：抽查 {len(sample_indices)} 个子节点，全部正确")
-        else:
-            print(f"  ❌ {level_name}：发现 {errors} 处错误！")
-
-
-# ── 入口 ──────────────────────────────────────────────────────
-
-if __name__ == '__main__':
-    import os
-
-    lineage_json  = "outputs/lineage.json"
-    genealogy_pkl = "outputs/genealogy.pkl"
-
-    genealogy = convert_lineage_json_to_genealogy_pkl(
-        lineage_json_path=lineage_json,
-        save_pkl_path=genealogy_pkl,
-        max_children=4,
-    )
-
-    verify_conversion(lineage_json, genealogy_pkl, n_sample=10)