feat: update script names and enhance iterative evaluation with table generation and plotting

BENCHMARKING.md

@@ -243,7 +243,7 @@ config/loss/ckpt recorded in the JSONs), and a no-op
 
 The mirror of A's ablation on the MONAI backbone: C trained with every
 hyperparameter at C's own default (its original study protocol from
-`scripts_monai/02_monai_train.py`, nothing nnU-derived) — BatchNorm
+`scripts_monai/07_train.py`, nothing nnU-derived) — BatchNorm
 `UNet(16→128)` with MONAI default init, LeakyReLU (C's own default act),
 per-sample 0.5CE+0.5mean-dice loss (WeightedDiceCELoss, no deep
 supervision), AdamW 3e-4 (wd 1e-4) + warmup/cosine, grad clip 5.0,
@@ -314,7 +314,7 @@ Artifacts: `results/benchmark_full/benchmark_full_C_legacy_std_100ep.json`
   0.4002 @100ep record; `--legacy` reverts). Checkpoint conventions: see the
   A and C ablation sections.
 - Study entrypoints match the record configs (2026-10-07): `scripts/07_train.py`
-  (A) was already A-native all-defaults; `scripts_monai/02_monai_train.py`
+  (A) was already A-native all-defaults; `scripts_monai/07_train.py`
   (C) now trains B-aligned by default (its record config) — `--legacy` reverts
   to the pre-2026-10-07 config. C checkpoints carry a `b_aligned` marker that
   03/04 read for the Z-score input; in C's B-aligned study mode pseudo rows

README.md

@@ -95,6 +95,10 @@ Run the full study:
 python scripts/09_run_iterative.py --rounds 4 --gpus 3
 ```
 
+Per-round test results are saved to `results/round{k}_test.json`, and
+`results/iterative_table.jsonl` + `results/iterative_dice.png` are rewritten
+after every round's evaluation.
+
 Individual stages are standalone, e.g.:
 
 ```bash
@@ -140,12 +144,12 @@ Dataset: `Dataset210_NTUH_T1C_PL`, single channel `t1c`, 2 classes
 
 | # | Script | Purpose |
 |---|---|---|
-| 01 | `01_nnu_prepare_dataset.py` | Build/rebuild the raw dataset (symlinked **native** `imagesTr`, `labelsTr` 0/1 + `dataset.json`) from a rows jsonl |
+| 01 | `01_prepare_dataset.py` | Build/rebuild the raw dataset (symlinked **native** `imagesTr`, `labelsTr` 0/1 + `dataset.json`) from a rows jsonl |
-| 02 | `02_nnu_plan_preprocess.py` | Plan + preprocess (`--clean`), then write subject-level `splits_final.json` |
+| 02 | `02_plan_preprocess.py` | Plan + preprocess (`--clean`), then write subject-level `splits_final.json` |
-| 03 | `03_nnu_train.py` | Train one round via `NTUHLPLTrainer` (DDP `-num_gpus`), optional warm start |
+| 06 | `06_pseudo_label.py` | Multi-GPU `nnUNetv2_predict` on native pool volumes + selection gates (physical mm) |
-| 04 | `04_nnu_pseudo_label.py` | Multi-GPU `nnUNetv2_predict` on native pool volumes + selection gates (physical mm) |
+| 07 | `07_train.py` | Train one round via `NTUHLPLTrainer` (DDP `-num_gpus`), optional warm start |
-| 05 | `05_nnu_eval_test.py` | Held-out test evaluation on native volumes (probability Dice@0.5 + hard-seg Dice) |
+| 08 | `08_eval.py` | Held-out test evaluation on native volumes (probability Dice@0.5 + hard-seg Dice) |
-| 06 | `06_nnu_run_iterative.py` | Orchestrates rounds 0…K, table + plot |
+| 09 | `09_run_iterative.py` | Orchestrates rounds 0…K, table + plot |
 | — | `trainers/ntuh_pl_trainer.py` | `NTUHLPLTrainer`: env-driven epochs/LR + full-weight warm start |
 | — | `nnu_common.py` | Shared paths, env, dataset/split helpers, native-grid selection + consistency |
 
@@ -210,25 +214,26 @@ across visits.
 Full study from the repo root:
 
 ```bash
-python scripts_nnu/06_nnu_run_iterative.py --rounds 4 --gpus 3
+python scripts_nnu/09_run_iterative.py --rounds 4 --gpus 3
 ```
 
 Defaults: baseline 250 epochs @ 1e-2; warm-started rounds 75 epochs @ 1e-3.
 Pseudo-label gates match Pipeline A (`--tau-pos 0.95`, `--min-cc-frac 0.2`,
 `--neg-frac 0.9`, volume p2–p98). Optional flags: `--no-tta`,
 `--no-neg-pseudo`, and the gate overrides. Outputs to `results/round{k}_test_nnu.json`,
-`results/iterative_table_nnu.jsonl`, `results/iterative_dice_nnu.png`; per-round
+`results/iterative_table_nnu.jsonl` (rewritten after every round's evaluation),
+`results/iterative_dice_nnu.png`; per-round
 checkpoints snapshotted to `runs_nnu/round{k}/best_nnu.pth`.
 
 Individual stages:
 
 ```bash
-python scripts_nnu/01_nnu_prepare_dataset.py --rows <rows.jsonl>
+python scripts_nnu/01_prepare_dataset.py --rows <rows.jsonl>
-python scripts_nnu/02_nnu_plan_preprocess.py --val data/manifests/split_val.jsonl
+python scripts_nnu/02_plan_preprocess.py --val data/manifests/split_val.jsonl
-python scripts_nnu/03_nnu_train.py --gpus 3 --epochs 250 --lr 1e-2
+python scripts_nnu/07_train.py --gpus 3 --epochs 250 --lr 1e-2
-python scripts_nnu/04_nnu_pseudo_label.py --pool data/manifests/unlabeled_pool.jsonl \
+python scripts_nnu/06_pseudo_label.py --pool data/manifests/unlabeled_pool.jsonl \
     --out data/pseudo_nnu/round1 --gpus 3
-python scripts_nnu/05_nnu_eval_test.py --rows data/manifests/split_test.jsonl \
+python scripts_nnu/08_eval.py --rows data/manifests/split_test.jsonl \
     --out results/round0_test_nnu.json --gpus 3
 ```
 
@@ -278,11 +283,11 @@ directly comparable to Pipelines A and B.
 
 | # | Script | Purpose |
 |---|---|---|
-| 01 | `01_monai_build_rows.py` | Assemble a round's training manifest (labeled + accepted pseudo rows, weighted, de-duped by key) |
+| 01 | `01_build_rows.py` | Assemble a round's training manifest (labeled + accepted pseudo rows, weighted, de-duped by key) |
-| 02 | `02_monai_train.py` | DDP training of the MONAI UNet for one round (torchrun), full-weight warm start (same config mode), auto-resume |
+| 06 | `06_pseudo_label.py` | Sliding-window + TTA prediction of the remaining pool, selection gates + consistency filter |
-| 03 | `03_monai_pseudo_label.py` | Sliding-window + TTA prediction of the remaining pool, selection gates + consistency filter |
+| 07 | `07_train.py` | DDP training of the MONAI UNet for one round (torchrun), full-weight warm start (same config mode), auto-resume |
-| 04 | `04_monai_eval_test.py` | Held-out test evaluation (Dice from probability map @0.5) |
+| 08 | `08_eval.py` | Held-out test evaluation (Dice from probability map @0.5) |
-| 05 | `05_monai_run_iterative.py` | Orchestrates rounds 0…K, table + plot |
+| 09 | `09_run_iterative.py` | Orchestrates rounds 0…K, table + plot |
 | — | `monai_common.py` | Network (legacy + B-aligned) / losses / transform chain / inference + distributed helpers |
 
 ### Running
@@ -290,7 +295,7 @@ directly comparable to Pipelines A and B.
 Full study from the repo root:
 
 ```bash
-python scripts_monai/05_monai_run_iterative.py --rounds 4 --gpus 3
+python scripts_monai/09_run_iterative.py --rounds 4 --gpus 3
 ```
 
 Defaults: B-aligned protocol (SGD 1e-2 + PolyLR, restarted per round); baseline
@@ -300,24 +305,25 @@ pre-2026-10-07 config (baseline 100 epochs @ 3e-4, warm-started rounds
 Pipelines A/B (`--tau-pos 0.95`, `--min-cc-frac 0.2`, `--neg-frac 0.9`, volume
 p2–p98). Optional flags: `--no-tta`, `--no-neg-pseudo`, `--legacy`, and the gate
 overrides. Outputs to
-`results/round{k}_test_monai.json`, `results/iterative_table_monai.jsonl`,
+`results/round{k}_test_monai.json`, `results/iterative_table_monai.jsonl`
+(rewritten after every round's evaluation),
 `results/iterative_dice_monai.png`; per-round checkpoints in
 `runs_monai/round{k}/best.pt`.
 
 Individual stages:
 
 ```bash
-python scripts_monai/01_monai_build_rows.py --train data/manifests/split_train.jsonl \
+python scripts_monai/01_build_rows.py --train data/manifests/split_train.jsonl \
     --accepted data/pseudo_monai/round1/accepted.jsonl \
     --out data/pseudo_monai/round1_rows.jsonl
-torchrun --standalone --nproc_per_node 3 scripts_monai/02_monai_train.py \
+torchrun --standalone --nproc_per_node 3 scripts_monai/07_train.py \
     --rows data/pseudo_monai/round1_rows.jsonl --val data/manifests/split_val.jsonl \
     --epochs 30 --batch 3 --ckpt-dir runs_monai/round1 \
     --pretrained runs_monai/round0/best.pt
-torchrun --standalone --nproc_per_node 3 scripts_monai/03_monai_pseudo_label.py \
+torchrun --standalone --nproc_per_node 3 scripts_monai/06_pseudo_label.py \
     --ckpt runs_monai/round1/best.pt --pool data/manifests/unlabeled_pool.jsonl \
     --out data/pseudo_monai/round2 --already data/pseudo_monai/added_keys.jsonl
-torchrun --standalone --nproc_per_node 3 scripts_monai/04_monai_eval_test.py \
+torchrun --standalone --nproc_per_node 3 scripts_monai/08_eval.py \
     --rows data/manifests/split_test.jsonl --ckpt runs_monai/round1/best.pt \
     --out results/round1_test_monai.json
 ```

scripts/09_run_iterative.py

@@ -39,6 +39,54 @@ def torchrun(nproc, script, extra):
     return f"{tr} --standalone --nproc_per_node {nproc} {script} {extra}"
 
 
+def build_table(args):
+    """Scan per-round test results, rewrite results/iterative_table.jsonl + plot.
+
+    Called after every round's evaluation so each round's performance is saved
+    incrementally (survives interrupted runs).
+    """
+    results_dir = f"{ROOT}/results"
+    table = []
+    for k in range(args.rounds + 1):
+        f = os.path.join(results_dir, f"round{k}_test.json")
+        if not os.path.exists(f):
+            continue
+        r = json.load(open(f))
+        row = {"round": k, "test_dice": round(r["dice"], 4), "n_test": r["n"],
+               "ckpt_epoch": r.get("epoch")}
+        pf = os.path.join(DATA, "pseudo", f"round{k}", "summary.json") if k > 0 else None
+        if pf and os.path.exists(pf):
+            s = json.load(open(pf))
+            row.update({"n_pos": s["n_pos"], "n_neg": s["n_neg"],
+                        "n_rej_cons": s["n_rejected_consistency"],
+                        "pos_vol_med_mm3": s["pos_vol_mm3"]["med"]})
+        sf = os.path.join(ROOT, "runs", f"round{k}", "final.pt")
+        row["ckpt"] = sf if os.path.exists(sf) else ""
+        table.append(row)
+    save_jsonl(table, os.path.join(results_dir, "iterative_table.jsonl"))
+    if not table:
+        return table
+    print(json.dumps(table, indent=1))
+    try:
+        import matplotlib
+        matplotlib.use("Agg")
+        import matplotlib.pyplot as plt
+        rs = [t["round"] for t in table]
+        ds = [t["test_dice"] for t in table]
+        plt.figure(figsize=(6, 4))
+        plt.plot(rs, ds, "o-")
+        plt.xlabel("pseudo-labeling round")
+        plt.ylabel("holdout tumor Dice")
+        for x, y in zip(rs, ds):
+            plt.annotate(f"{y:.3f}", (x, y), textcoords="offset points", xytext=(0, 8), fontsize=8)
+        plt.grid(alpha=0.3)
+        plt.tight_layout()
+        plt.savefig(os.path.join(results_dir, "iterative_dice.png"), dpi=150)
+    except Exception as e:  # noqa
+        print("plot failed:", repr(e))
+    return table
+
+
 def main():
     ap = argparse.ArgumentParser()
     ap.add_argument("--rounds", type=int, default=4)
@@ -67,6 +115,7 @@ def main():
                  f"--rows {train_f} --val {val_f} --epochs {args.base_epochs} --lr {args.lr} "
                  f"--batch {args.batch} --ckpt-dir runs/round0"), f"{main_log}")
     run(f"python scripts/08_eval.py --rows {test_f} --ckpt runs/round0/best.pt", f"{main_log}")
+    build_table(args)
 
     for k in range(1, args.rounds + 1):
         pdir = os.path.join(DATA, "pseudo/round" + str(k))
@@ -91,42 +140,10 @@ def main():
                      f"--lr {args.round_lr} --batch {args.batch} --resume runs/round{k-1}/best.pt "
                      f"--ckpt-dir runs/round{k} --val-every 1"), f"{main_log}")
         run(f"python scripts/08_eval.py --rows {test_f} --ckpt runs/round{k}/best.pt", f"{main_log}")
+        build_table(args)
 
     # ---- report ----
-    table = []
+    build_table(args)
-    for k in list(range(args.rounds + 1)):
-        f = os.path.join(results_dir, f"round{k}_test.json")
-        if os.path.exists(f):
-            r = json.load(open(f))
-            table.append({"round": k, "test_dice": round(r["dice"], 4), "n_test": r["n"],
-                          "ckpt_epoch": r.get("epoch")})
-            pf = os.path.join(DATA, f"pseudo/round{k}/summary.json") if k > 0 else None
-            if pf and os.path.exists(pf):
-                s = json.load(open(pf))
-                table[-1].update({"n_pos": s["n_pos"], "n_neg": s["n_neg"],
-                                  "n_rej_cons": s["n_rejected_consistency"],
-                                  "pos_vol_med_mm3": s["pos_vol_mm3"]["med"]})
-        sf = os.path.join(ROOT, f"runs/round{k}/final.pt")
-        table[-1]["ckpt"] = sf if os.path.exists(sf) else ""
-    save_jsonl(table, os.path.join(results_dir, "iterative_table.jsonl"))
-    print(json.dumps(table, indent=1))
-    try:
-        import matplotlib
-        matplotlib.use("Agg")
-        import matplotlib.pyplot as plt
-        rs = [t["round"] for t in table]
-        ds = [t["test_dice"] for t in table]
-        plt.figure(figsize=(6, 4))
-        plt.plot(rs, ds, "o-")
-        plt.xlabel("pseudo-labeling round")
-        plt.ylabel("holdout tumor Dice")
-        for x, y in zip(rs, ds):
-            plt.annotate(f"{y:.3f}", (x, y), textcoords="offset points", xytext=(0, 8), fontsize=8)
-        plt.grid(alpha=0.3)
-        plt.tight_layout()
-        plt.savefig(os.path.join(results_dir, "iterative_dice.png"), dpi=150)
-    except Exception as e:  # noqa
-        print("plot failed:", repr(e))
 
 
 if __name__ == "__main__":

scripts_monai/01_build_rows.py

@@ -3,12 +3,12 @@
 base = labeled patient-level train split (w=1.0) — the same train/val usage as
 Pipeline A: the val split is never trained on (it is the internal-validation
 set). Accepted pseudo-label rows (pos + neg) are appended with w=<pseudo-weight>
-(negative rows point at the zero mask written by stage 03), de-duplicated by
+(negative rows point at the zero mask written by stage 06), de-duplicated by
 key (first occurrence wins). Every row must point at existing pimg/plabel
 niftis.
 
 Usage:
-    python scripts_monai/01_monai_build_rows.py \
+    python scripts_monai/01_build_rows.py \
         --train data/manifests/split_train.jsonl \
         --accepted data/pseudo_monai/round1/accepted.jsonl \
         --pseudo-weight 0.3 --out data/pseudo_monai/round1_rows.jsonl

scripts_monai/06_pseudo_label.py

@@ -1,6 +1,6 @@
 """Pseudo-label the unlabeled pool with the MONAI round model (torchrun; rank = GPU).
 
-torchrun --standalone --nproc_per_node 3 scripts_monai/03_monai_pseudo_label.py \
+torchrun --standalone --nproc_per_node 3 scripts_monai/06_pseudo_label.py \
     --ckpt runs_monai/round0/best.pt --pool data/manifests/unlabeled_pool.jsonl \
     --out data/pseudo_monai/round1 [--already data/pseudo_monai/added_keys.jsonl]
 
@@ -18,8 +18,8 @@ mask for neg), <out>/accepted.jsonl, <out>/summary.json. Per-rank shards are
 flushed to <out>/part{rank}.jsonl so interrupted runs resume.
 
 Usage:
-    python scripts_monai/03_monai_pseudo_label.py <same flags, single GPU>
+    python scripts_monai/06_pseudo_label.py <same flags, single GPU>
-    torchrun --standalone --nproc_per_node 3 scripts_monai/03_monai_pseudo_label.py ...
+    torchrun --standalone --nproc_per_node 3 scripts_monai/06_pseudo_label.py ...
 """
 import argparse
 import os

scripts_monai/07_train.py

@@ -1,6 +1,6 @@
 """MONAI DDP training entrypoint for one pseudo-labeling round (torchrun).
 
-torchrun --standalone --nproc_per_node 3 scripts_monai/02_monai_train.py \
+torchrun --standalone --nproc_per_node 3 scripts_monai/07_train.py \
     --rows data/pseudo_monai/round0_rows.jsonl --val data/manifests/split_val.jsonl \
     --epochs 100 --batch 3 --ckpt-dir runs_monai/round0 \
     [--pretrained runs_monai/round0/best.pt]

scripts_monai/08_eval.py

@@ -1,6 +1,6 @@
 """Held-out test evaluation of a MONAI round model (torchrun; rank = GPU).
 
-torchrun --standalone --nproc_per_node 3 scripts_monai/04_monai_eval_test.py \
+torchrun --standalone --nproc_per_node 3 scripts_monai/08_eval.py \
     --rows data/manifests/split_test.jsonl --ckpt runs_monai/round0/best.pt \
     --out results/round0_test_monai.json [--no-tta]
 
@@ -12,8 +12,8 @@ Per-rank results are appended to <out>_part{rank}.jsonl for resume; rank 0
 merges them into <out> (json) + <out>_per_row.jsonl.
 
 Usage:
-    python scripts_monai/04_monai_eval_test.py <same flags, single GPU>
+    python scripts_monai/08_eval.py <same flags, single GPU>
-    torchrun --standalone --nproc_per_node 3 scripts_monai/04_monai_eval_test.py ...
+    torchrun --standalone --nproc_per_node 3 scripts_monai/08_eval.py ...
 """
 import argparse
 import os

scripts_monai/09_run_iterative.py

@@ -1,5 +1,5 @@
 """Orchestrates the MONAI iterative pseudo-labeling study (parallel to
-scripts/09 and scripts_nnu/06).
+scripts/09 and scripts_nnu/09).
 
 Round 0: train the MONAI UNet on the labeled patient-level train split (internal
           validation = held-out patient-level val split, subject-disjoint).
@@ -15,7 +15,7 @@ Config: rounds train with the B-aligned protocol (C's record benchmark config,
         --legacy (pre-2026-10-07 config: AdamW at --base-lr/--round-lr +
         per-sample weighted Dice+CE with pseudo-row down-weighting).
 
-Usage: python scripts_monai/05_monai_run_iterative.py [--rounds 4] [--gpus 3] [--legacy]
+Usage: python scripts_monai/09_run_iterative.py [--rounds 4] [--gpus 3] [--legacy]
 """
 import os
 import sys
@@ -101,7 +101,7 @@ def main():
             extra += " --no-neg"
         for f in accepted:
             extra += f" --accepted {f}"
-        py("01_monai_build_rows.py", extra)
+        py("01_build_rows.py", extra)
 
     def train(round, rows_f, epochs, lr, warmstart, val_every):
         extra = (f"--rows {rows_f} --val {val_f} --epochs {epochs} --lr {lr} "
@@ -111,7 +111,7 @@ def main():
             extra += f" --pretrained {warmstart}"
         if args.legacy:
             extra += " --legacy"
-        trrun("02_monai_train.py", extra, os.path.join(logdir, f"train_r{round}.log"))
+        trrun("07_train.py", extra, os.path.join(logdir, f"train_r{round}.log"))
 
     def predict_pool(round, ckpt, added=None):
         out = os.path.join(pdir_root, f"round{round}")
@@ -123,7 +123,7 @@ def main():
             extra += f" --already {added}"
         if args.no_tta:
             extra += " --no-tta"
-        trrun("03_monai_pseudo_label.py", extra, os.path.join(logdir, f"pseudo_r{round}.log"))
+        trrun("06_pseudo_label.py", extra, os.path.join(logdir, f"pseudo_r{round}.log"))
         return os.path.join(out, "accepted.jsonl")
 
     def eval_test(round, ckpt):
@@ -131,29 +131,15 @@ def main():
         extra = f"--rows {test_f} --ckpt {ckpt} --out {out}"
         if args.no_tta:
             extra += " --no-tta"
-        trrun("04_monai_eval_test.py", extra, os.path.join(logdir, f"eval_r{round}.log"))
+        trrun("08_eval.py", extra, os.path.join(logdir, f"eval_r{round}.log"))
         return out
 
     def best(round):
         return os.path.join(ROOT, "runs_monai", f"round{round}", "best.pt")
 
-    # ---- round 0: baseline ----
+    def build_table():
-    rows0 = os.path.join(pdir_root, "round0_rows.jsonl")
+        # Rewrite the per-round table + plot; called after every round's eval so
-    build_rows(rows0, [])
+        # each round's performance is saved incrementally (survives interruptions).
-    train(0, rows0, args.base_epochs, args.base_lr, None, args.val_every)
-    accepted_f = [predict_pool(1, best(0))]
-    save_jsonl(load_jsonl(accepted_f[0]), added_f)
-    eval_test(0, best(0))
-
-    for k in range(1, args.rounds + 1):
-        rows_f = os.path.join(pdir_root, f"round{k}_rows.jsonl")
-        build_rows(rows_f, accepted_f, no_neg=args.no_neg_pseudo)
-        train(k, rows_f, args.round_epochs, args.round_lr, best(k - 1), args.val_every)
-        accepted_f.append(predict_pool(k + 1, best(k), added_f))
-        save_jsonl([r for f in accepted_f for r in load_jsonl(f)], added_f)
-        eval_test(k, best(k))
-
-    # ---- report (same layout as scripts/09 and scripts_nnu/06) ----
         table = []
         for k in range(args.rounds + 1):
             resf = os.path.join(results_dir, f"round{k}_test_monai.json")
@@ -169,6 +155,8 @@ def main():
                             "pos_vol_med_mm3": s["pos_vol_mm3"]["med"]})
             table.append(row)
         save_jsonl(table, os.path.join(results_dir, "iterative_table_monai.jsonl"))
+        if not table:
+            return table
         print(json.dumps(table, indent=1))
         try:
             import matplotlib
@@ -187,6 +175,28 @@ def main():
             plt.savefig(os.path.join(results_dir, "iterative_dice_monai.png"), dpi=150)
         except Exception as e:  # noqa
             print("plot failed:", repr(e))
+        return table
+
+    # ---- round 0: baseline ----
+    rows0 = os.path.join(pdir_root, "round0_rows.jsonl")
+    build_rows(rows0, [])
+    train(0, rows0, args.base_epochs, args.base_lr, None, args.val_every)
+    accepted_f = [predict_pool(1, best(0))]
+    save_jsonl(load_jsonl(accepted_f[0]), added_f)
+    eval_test(0, best(0))
+    build_table()
+
+    for k in range(1, args.rounds + 1):
+        rows_f = os.path.join(pdir_root, f"round{k}_rows.jsonl")
+        build_rows(rows_f, accepted_f, no_neg=args.no_neg_pseudo)
+        train(k, rows_f, args.round_epochs, args.round_lr, best(k - 1), args.val_every)
+        accepted_f.append(predict_pool(k + 1, best(k), added_f))
+        save_jsonl([r for f in accepted_f for r in load_jsonl(f)], added_f)
+        eval_test(k, best(k))
+        build_table()
+
+    # ---- report (same layout as scripts/09 and scripts_nnu/09) ----
+    build_table()
 
 
 if __name__ == "__main__":

scripts_nnu/01_prepare_dataset.py

@@ -7,7 +7,7 @@ preprocessing; imagesTr entries are symlinks to the native volumes, labelsTr
 entries are symlinks when already 0/1 on the image grid, else nearest-warped
 + binarized copies (label fix-up only).
 
-Usage: python scripts_nnu/01_nnu_prepare_dataset.py --rows <rows.jsonl>
+Usage: python scripts_nnu/01_prepare_dataset.py --rows <rows.jsonl>
 """
 import argparse
 import os

scripts_nnu/02_plan_preprocess.py

@@ -2,7 +2,7 @@
 subject-level splits_final.json: val = split_val cases, train = the rest
 (all base + pseudo cases), same 5-fold content so -f 0 is the one we use.
 
-Usage: python scripts_nnu/02_nnu_plan_preprocess.py --val data/manifests/split_val.jsonl [--npp 8]
+Usage: python scripts_nnu/02_plan_preprocess.py --val data/manifests/split_val.jsonl [--npp 8]
 """
 import argparse
 import os
@@ -23,7 +23,7 @@ def main():
     run(plan_preprocess_cmd(args.npp), log, env=nnu_env())
     val_keys = [r["key"] for r in load_jsonl(args.val)]
     all_keys = dataset_case_keys()
-    assert all_keys, "raw dataset is empty; run 01_nnu_prepare_dataset.py first"
+    assert all_keys, "raw dataset is empty; run 01_prepare_dataset.py first"
     assert set(val_keys) <= set(all_keys), f"val keys not in dataset: {set(val_keys) - set(all_keys)}"
     n_tr, n_va = write_splits(all_keys, val_keys)
     print(f"[nnu:plan] splits_final.json written to {preproc_ds()}: train={n_tr} val={n_va}", flush=True)

scripts_nnu/06_pseudo_label.py

@@ -17,7 +17,7 @@ Writes <out>/rows.jsonl, <out>/<key>_label.nii.gz (native grid), <out>/accepted.
    <out>/summary.json.
 
 Usage:
-  python scripts_nnu/04_nnu_pseudo_label.py --pool data/manifests/unlabeled_pool.jsonl \
+  python scripts_nnu/06_pseudo_label.py --pool data/manifests/unlabeled_pool.jsonl \
       --out data/pseudo_nnu/round1 --gpus 3
 """
 import argparse

scripts_nnu/07_train.py

@@ -4,7 +4,7 @@ Epochs / initial lr / warm-start checkpoint are passed to NTUHLPLTrainer via
 NNU_PL_* env vars (see scripts_nnu/trainers/ntuh_pl_trainer.py).
 
 Usage:
-  nnUNetv2_train wrapper: python scripts_nnu/03_nnu_train.py \
+  nnUNetv2_train wrapper: python scripts_nnu/07_train.py \
       --gpus 3 --epochs 250 --lr 1e-2 [--warmstart runs_nnu/round0/best_nnu.pth]
 """
 import argparse

scripts_nnu/08_eval.py

@@ -10,7 +10,7 @@ grid, so we score directly there:
 Ground truth is the row's native label.
 Writes: <out> json + <out>.json -> per_row jsonl alongside.
 
-Usage: python scripts_nnu/05_nnu_eval_test.py --rows data/manifests/split_test.jsonl \
+Usage: python scripts_nnu/08_eval.py --rows data/manifests/split_test.jsonl \
     --out results/round0_test_nnu.json --gpus 3
 """
 import argparse

scripts_nnu/09_run_iterative.py

@@ -8,7 +8,7 @@ Round k: dataset grows with accepted pseudo-labels (pos + neg) from rounds 1..k;
           unlabeled pool (gates identical to scripts/06) and evaluate the model on
           the held-out patient-level test split.
 
-Usage: python scripts_nnu/06_nnu_run_iterative.py [--rounds 4] [--gpus 3]
+Usage: python scripts_nnu/09_run_iterative.py [--rounds 4] [--gpus 3]
 """
 import os
 import sys
@@ -71,16 +71,16 @@ def main():
     def prepare(rows):
         f = os.path.join(pdir_root, "current_rows.jsonl")
         save_jsonl(rows, f)
-        py("01_nnu_prepare_dataset.py", f"--rows {f}")
+        py("01_prepare_dataset.py", f"--rows {f}")
 
     def plan():
-        py("02_nnu_plan_preprocess.py", f"--val {val_f}")
+        py("02_plan_preprocess.py", f"--val {val_f}")
 
     def train(round, epochs, lr, warmstart):
         extra = f"--gpus {args.gpus} --epochs {epochs} --lr {lr} --log {os.path.join(logdir, f'nnu_train_r{round}.log')}"
         if warmstart:
             extra += f" --warmstart {warmstart}"
-        py("03_nnu_train.py", extra)
+        py("07_train.py", extra)
         dst = os.path.join(ROOT, "runs_nnu", f"round{round}", "best_nnu.pth")
         os.makedirs(os.path.dirname(dst), exist_ok=True)
         shutil.copy2(best_ckpt(), dst)
@@ -95,40 +95,19 @@ def main():
             extra += " --no-tta"
         if added:
             extra += f" --already {added_f}"
-        py("04_nnu_pseudo_label.py", extra)
+        py("06_pseudo_label.py", extra)
         return os.path.join(pdir_root, f"round{round}", "accepted.jsonl")
 
     def eval_test(round, out_name):
         extra = (f"--rows {test_f} --out {results_dir}/{out_name} --gpus {args.gpus}")
         if args.no_tta:
             extra += " --no-tta"
-        py("05_nnu_eval_test.py", extra)
+        py("08_eval.py", extra)
         return os.path.join(results_dir, out_name)
 
-    # ---- round 0: baseline ----
+    def build_table():
-    prepare(base)
+        # Rewrite the per-round table + plot; called after every round's eval so
-    plan()
+        # each round's performance is saved incrementally (survives interruptions).
-    train(0, args.base_epochs, args.base_lr, None)
-    accepted_f = [predict_pool(1, None)]
-    save_jsonl(load_jsonl(accepted_f[0]), added_f)
-    eval_test(0, "round0_test_nnu.json")
-
-    for k in range(1, args.rounds + 1):
-        rows = list(base)
-        for f in accepted_f:
-            for r in load_jsonl(f):
-                if args.no_neg_pseudo and r["role"] == "neg":
-                    continue
-                if all(r["key"] != x["key"] for x in rows):
-                    rows.append(r)
-        prepare(rows)
-        plan()
-        train(k, args.round_epochs, args.round_lr, os.path.join(ROOT, "runs_nnu", f"round{k-1}", "best_nnu.pth"))
-        accepted_f.append(predict_pool(k + 1, added_f))
-        save_jsonl([r for f in accepted_f for r in load_jsonl(f)], added_f)
-        eval_test(k, f"round{k}_test_nnu.json")
-
-    # ---- report (same layout as scripts/09) ----
         table = []
         for k in range(args.rounds + 1):
             resf = os.path.join(results_dir, f"round{k}_test_nnu.json")
@@ -145,6 +124,8 @@ def main():
                             "pos_vol_med_mm3": s["pos_vol_mm3"]["med"]})
             table.append(row)
         save_jsonl(table, os.path.join(results_dir, "iterative_table_nnu.jsonl"))
+        if not table:
+            return table
         print(json.dumps(table, indent=1))
         try:
             import matplotlib
@@ -163,6 +144,35 @@ def main():
             plt.savefig(os.path.join(results_dir, "iterative_dice_nnu.png"), dpi=150)
         except Exception as e:  # noqa
             print("plot failed:", repr(e))
+        return table
+
+    # ---- round 0: baseline ----
+    prepare(base)
+    plan()
+    train(0, args.base_epochs, args.base_lr, None)
+    accepted_f = [predict_pool(1, None)]
+    save_jsonl(load_jsonl(accepted_f[0]), added_f)
+    eval_test(0, "round0_test_nnu.json")
+    build_table()
+
+    for k in range(1, args.rounds + 1):
+        rows = list(base)
+        for f in accepted_f:
+            for r in load_jsonl(f):
+                if args.no_neg_pseudo and r["role"] == "neg":
+                    continue
+                if all(r["key"] != x["key"] for x in rows):
+                    rows.append(r)
+        prepare(rows)
+        plan()
+        train(k, args.round_epochs, args.round_lr, os.path.join(ROOT, "runs_nnu", f"round{k-1}", "best_nnu.pth"))
+        accepted_f.append(predict_pool(k + 1, added_f))
+        save_jsonl([r for f in accepted_f for r in load_jsonl(f)], added_f)
+        eval_test(k, f"round{k}_test_nnu.json")
+        build_table()
+
+    # ---- report (same layout as scripts/09) ----
+    build_table()
 
 
 if __name__ == "__main__":