总览 · ← 返回运行 20261001-004247-search-t1-spark

节点 n4 终选程序?按该运行锁定的规则最终选出的程序；可能是候选节点，也可能由护栏回退到基线。在终选来历上

父组成 heart_reweight 不变，叠加"按类型沿单快照增殖分化轴的稀疏保形平移"(β=0.3)：每类型 d_c=低增殖三分位均值−高增殖三分位均值(早→晚)，只对该细胞已非零基因做 clip(x+β·d_c,0)，保留稀疏模式护住 covariation。

方法说明?节点程序自带的 METHOD.md：这个程序做了什么、为什么。

来自 programs.git dad83b8b31:solution/METHOD.md

调研员的计划

代码改动?这个节点的程序和父节点程序的逐行差别：绿色是新增，红色是删除。

对比：父节点版本 d8e3904ea4。改动的文件：solution/METHOD.md +26 −0、solution/run.py +108 −6

diff --git a/solution/METHOD.md b/solution/METHOD.mdnew file mode 100644index 0000000..9d8741a--- /dev/null+++ b/solution/METHOD.md@@ -0,0 +1,26 @@+父组成 heart_reweight 不变，叠加"按类型沿单快照增殖分化轴的稀疏保形平移"(β=0.3)：每类型 d_c=低增殖三分位均值−高增殖三分位均值(早→晚)，只对该细胞已非零基因做 clip(x+β·d_c,0)，保留稀疏模式护住 covariation。++## 方法++- 父节点(seed heart_jcf_peri, proxy 55.97)只重加权"哪些类型在场"并复制真实细胞，最弱组是 de_recovery 53.06，且从不把每个类型推向更分化状态。单阶段 proxy 上真实两阶段差为空(pseudobulk_shift 退化成 copy_last)，唯一能算的表达杠杆是单快照分化轴。+- 逐类型方向 d_c：用通用细胞周期基因(S/G2M: Mki67/Top2a/Cdk1/Ccnb1/Ccna2/Birc5/Pcna/Rrm2/Tyms/Mcm2-7/Cdk2/Ccnd1/Ccne1/Rrm1/Cdc20/Ube2c/Plk1/Aurkb/Bub1/Kif11/Ccnb2/Cdca7/Gins2/Chaf1b/Slbp，取 panel 中存在者)算每细胞增殖分 p_i=log1p 均值；类型内按 p 排序，低三分位=late(成熟)、高三分位=early(祖)，d_c=mean(late)−mean(early)，指向时间前进。+- 平移：对 heart_reweight 抽到的每类型细胞，x_stored ← clip(x_stored + β·d_c, 0)，**只改已存储的非零元素**，零模式不变。逐类型常数平移与行抽样可交换，故复制父节点 RNG(reweight.largest_remainder/type_weights/take 同序)后按类型平移，抽到的细胞集合与父完全一致；β=0 逐位复现父节点。++## 关键参数++- β=0.30(平台区 0.25–0.35，见下)；MIN_DIR_CELLS=50(细胞数<50 的类型不平移)；MIN_CC_GENES=3(细胞周期基因<3 则退回纯 heart_reweight)；lam 收缩=0(未用，见下)；N_CELLS=4000(经 target_n_cells 夹到 [min,max])。++## 验证过++- β=0 与父节点预测逐位相同(maxdiff=0)，vec-check ok，run 用时~2s、峰值内存远低于 28GB。+- 关键发现：稠密平移(lam=0，改全部基因)使 nnz/cell 3971→11109，covariation 崩(54.97→46.55)，净分反降(55.79)；**只改非零元素**保稀疏(nnz~4096)，covariation 保住(55.48)，净分升到 57.4。基因级 SD 收缩(lam)在单细胞噪声下过强(lam=1 把方向压到~0、无效)，故不用。+- β 扫描(nz_only,lam=0,seed0)：0→55.97, 0.15→57.08, 0.25→57.40, 0.35→57.43, 0.5→57.07；峰平台 0.25–0.35。+- β=0.3 三 seed：57.39/57.55/57.60(均值 57.51, std~0.1)；父 β=0 三 seed：55.97/56.32/55.90(均值 56.06, std~0.2)。增益 +1.45，远超噪声。分组均值：cell_state 56.9→60.7、direction 58.6→60.6、covariation 55.0→55.4、de_recovery 53.1→52.4(小降)。+- 增益主要来自 cell_state(mmd_u 0.0126→0.0107，细胞分布更贴近 E9.5)与 direction，covariation 因保稀疏而未受损。de_recovery 反而小降：单快照增殖轴不是精确的 DE 方向。++## 没验证 / 风险++- 迁移：方法只用 last 阶段，proxy(E8.5)与 final(E8.5+E9.5→读 E9.5)跑同段代码、含义相同，无分叉；但 final 的真实 E10.5 上增殖轴是否与时间轴同向、β=0.3 是否仍是最优，未测(替代评测只有一个输入阶段，测不到两阶段趋势)。final 未加两阶段 delta 融合增强。+- 增殖定向对"分化后仍增殖"的类型可能反向；已逐类型独立估计，但未做方向可信度门控(风险 2 未处理)。+- de_recovery 小降；若后续要抬 de_recovery，需换更贴近真实 DE 的方向(如允许外部注释/两阶段差)，而非增殖轴。+- 合规：逐类型常数平移不改坐标尺度、不旋转；用通用细胞周期基因而非保留阶段标记；标签取自 labels_of，名字随阶段自然迁移，未写死 E10.5/禁窗名字、比例或表达。diff --git a/solution/run.py b/solution/run.pyindex a752b21..160a411 100644--- a/solution/run.py+++ b/solution/run.py@@ -1,16 +1,39 @@ #!/usr/bin/env python3-"""heart_jcf_peri: run2's T1 winner. Reweight the latest input stage by cell type.+"""heart_maturation: parent heart_jcf_peri composition + per-type maturation shift. -Neural Tube dropped; heart types (CM, endothelium/endocardium, SHF/PHM, JCF,-pericardium, proepicardium) x1.6; surface ectoderm, EXEM, paraxial mesoderm-x0.25; everything else x1. 4000 real cells, no expression shift. Proxy reads-E8.5, final reads E9.5; the type set already names both stages' labels.+Parent (seed heart_jcf_peri, proxy 55.97) only rewights which cell types are+present and copies real cells; its weakest group was de_recovery 53.06, and it+never moves each type toward its more-differentiated state. On a single-stage+proxy the true two-stage temporal delta is empty, so the only expression lever+is a single-snapshot maturation axis.++For each cell type we build a direction d_c = mean(low-proliferation tercile)+- mean(high-proliferation tercile) using generic cell-cycle genes: high+proliferation = early/progenitor, low = mature/late, so d_c points early->late+(time forward). We then shift each sampled cell by beta*d_c, but ONLY on that+cell's already-nonzero genes (sparsity-preserving). Shifting only stored entries+keeps the exact sparsity pattern, which protects the covariation/variogram group+(a dense per-gene shift densifies the matrix and crashes covariation 55 -> 47).++The shift is a per-type constant, so it commutes with heart_reweight's row+sampling: we replicate the parent's RNG exactly and shift each type's sampled+rows, giving the identical cell set as the parent with expression moved along+d_c. beta=0 reproduces the parent byte-for-byte.++Proxy beta=0.3 (3 seeds): board 57.51 vs parent 56.06 (+1.45), driven by+cell_state 56.9->60.5 and direction 58.6->60.5, covariation held ~55.4.+Only the last input stage is used, so the same code runs on the final view+(E8.5,E9.5 -> reads E9.5); no stage names are hardcoded. """  from __future__ import annotations  import argparse +import numpy as np+from scipy import sparse++from src.task1_temporal import reweight as RW from src.task1_temporal.reweight import heart_reweight from src.task1_temporal.view_io import (     inputs_by_time,@@ -23,6 +46,78 @@ from src.task1_temporal.view_io import ( )  N_CELLS = 4000+BETA = 0.30            # maturation step in log space (flat optimum 0.25-0.35)+MIN_DIR_CELLS = 50     # types with fewer cells get no shift (unstable direction)+MIN_CC_GENES = 3       # need enough cell-cycle genes to define a proliferation axis++# generic cell-cycle (S/G2M) genes; not stage- or lineage-specific markers+CELL_CYCLE = [+    'Mki67', 'Top2a', 'Cdk1', 'Ccnb1', 'Ccna2', 'Birc5', 'Pcna', 'Rrm2', 'Tyms',+    'Mcm2', 'Mcm3', 'Mcm4', 'Mcm5', 'Mcm6', 'Mcm7', 'Cdk2', 'Ccnd1', 'Ccne1',+    'Rrm1', 'Cdc20', 'Ube2c', 'Plk1', 'Aurkb', 'Bub1', 'Kif11', 'Ccnb2', 'Cdca7',+    'Gins2', 'Chaf1b', 'Slbp',+]+++def proliferation(X, genes: list[str]) -> np.ndarray | None:+    """Per-cell mean log1p over the cell-cycle genes present in the panel."""+    gset = set(genes)+    idx = [genes.index(g) for g in CELL_CYCLE if g in gset]+    if len(idx) < MIN_CC_GENES:+        return None+    sub = X[:, idx]+    sub = sub.toarray() if sparse.issparse(sub) else np.asarray(sub)+    return sub.mean(axis=1)+++def type_directions(X, labels, genes: list[str]) -> dict[str, np.ndarray]:+    """d_c = mean(low-prolif tercile) - mean(high-prolif tercile), early->late."""+    p = proliferation(X, genes)+    dirs: dict[str, np.ndarray] = {}+    if p is None:+        return dirs+    for t in np.unique(labels):+        idx = np.flatnonzero(labels == t)+        if idx.size < MIN_DIR_CELLS:+            continue+        order = idx[np.argsort(p[idx])]          # ascending proliferation+        k = max(1, order.size // 3)+        late = order[:k]                          # low proliferation = mature+        early = order[-k:]                        # high proliferation = progenitor+        m_late = np.asarray(X[late].mean(axis=0)).ravel()+        m_early = np.asarray(X[early].mean(axis=0)).ravel()+        dirs[str(t)] = (m_late - m_early).astype(np.float32)+    return dirs+++def reweight_maturation(X, labels, n_cells, dirs, beta, seed):+    """Parent heart_reweight sampling (identical RNG) + sparsity-preserving shift.++    Each type's sampled rows get x_stored <- clip(x_stored + beta*d_c, 0); the+    zero pattern is untouched, so covariation is preserved. beta=0 (or a type+    with no direction) copies rows unchanged, matching the parent exactly.+    """+    rng = np.random.default_rng(seed)+    types = [str(t) for t in np.unique(labels) if str(t) not in RW.DROP_TYPES]+    counts = np.array([(labels == t).sum() for t in types], dtype=np.float64)+    alloc = RW.largest_remainder(+        counts * RW.type_weights(types, RW.HEART_WEIGHT, RW.EDGE_WEIGHT), n_cells+    )+    blocks = []+    for t, n in zip(types, alloc):+        if n <= 0:+            continue+        pool = np.flatnonzero(labels == t)+        choice = rng.choice(pool, size=int(n), replace=pool.size < int(n))+        rows = X[choice]+        if beta > 0 and t in dirs:+            rows = rows.copy().tocsr()+            d = dirs[t]+            rows.data = np.clip(rows.data + beta * d[rows.indices], 0, None).astype(np.float32)+        blocks.append(rows)+    out = sparse.vstack(blocks, format="csr").astype(np.float32)+    out.eliminate_zeros()+    return out   def main() -> None:@@ -35,8 +130,15 @@ def main() -> None:     manifest = load_manifest(args.data)     genes = panel_genes(args.data, manifest)     last = read_stage(args.data, inputs_by_time(manifest)[-1], genes)+    labels = labels_of(last)     n = target_n_cells(manifest, N_CELLS)-    X = heart_reweight(last.X, labels_of(last), n_cells=n, seed=args.seed)++    dirs = type_directions(last.X, labels, genes) if BETA > 0 else {}+    if dirs:+        X = reweight_maturation(last.X, labels, n, dirs, BETA, args.seed)+    else:+        # no usable proliferation axis: fall back to the exact parent behaviour+        X = heart_reweight(last.X, labels, n_cells=n, seed=args.seed)     write_prediction(X, genes, args.out, seed=args.seed)  

调研来源?调研员查到并用到的知识条目和文献检索结果（只列标题和编号）。

用到的知识库条目

分析结果?分析员写的 ANALYSIS.json：改了什么、各组分数怎么变、假设是否成立、经验和下一步建议。

对话摘要?每个角色和大模型对话的统计：轮数、工具调用、用时、token 数和最后的回答摘录；原始记录只给路径。

只给统计和最后回答的摘录；完整对话请到原始记录位置里列出的文件看。

分析员

ANALYSIS.json written

工程师

Robust improvement confirmed. Baseline mean 56.06 (s0/s1/s2: 55.97/56.32/55.90), beta=0.3 nz_only mean 57.51 (57.39/57.55/57.60) → **+1.45, far above noise**. Now finalizing the solution.

调研员

PLAN.json written

审查员

REVIEW.json written