updated pipeline to fix sequence_id bug, added additional step to greedily assign clusters to splits

data/fireprotdb_splits_balanced_ddg.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69d2b51712ea537f5229d1194ee5e4eae34e3db610ec0cefd0d2282415b05a82
+size 116931745

data/fireprotdb_with_cluster_splits.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a7685586c8a203b6548f7379f99e3dfec24fd1c8b161df8f4d879da455a86b4
-size 109389319
+oid sha256:0231c790251fa9e54aa22beba469753840329fca0b10fc511ed9bce1ae4db99e
+size 117032379

data/subsets/mutation_binary/test.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:46e35999b00ac379c7508c0def79f52b800f1bdaed2bfe77ef36461688ccb185
-size 62766
+oid sha256:5fcbae17160d47abbebe881b6b9cd4e38fea3325344209b6e0f736c06206fa64
+size 2399800

data/subsets/mutation_binary/train.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f2459ac8a7be1a454829874c5802ebbc4a7088f5c3597cfce791cfa8ef98240
-size 454356
+oid sha256:82b22ca971643cd76402ed24684565211c82ff1093fe2142ae1f29ccbaf4f963
+size 14543037

data/subsets/mutation_binary/validation.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72bb3dc7ba474e9ae11a1c7020d0b1af09b1f44b82cc5697c7b559af661d3dc6
-size 13100753
+oid sha256:0cb0fc05783975c2f54ba1b320ecb50e553da5576a08d1bc7161700875fccbbe
+size 2390616

data/subsets/mutation_ddg/test.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e4bdd7661b8d9b708fa5a7b125bd5803a0e0910421d70bbe2373cd4aa63a637
-size 63049
+oid sha256:cdc2c09cc26d4b86a5258b89fb09f08b619b40497ee61806b78d76b576c3c572
+size 2400677

data/subsets/mutation_ddg/train.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1314c1bf86e41315bd817b3a73ddb23b2e4bba20a7b581473e3fa7dd1d70cb56
-size 458469
+oid sha256:d6407dde0c4b5f36080ebca6eed5091c82484d17c619786f20250d127d5d71fd
+size 14548884

data/subsets/mutation_ddg/validation.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d99c5f983d01ce91294122f84c1e4622c7bad28b046abc4e803d344f2cf15742
-size 13099916
+oid sha256:8e77dc84cf0087f69d310bfb44a23565ca794f582da5cd4727d2dc681d756469
+size 2391573

data/subsets/mutation_dtm/test.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:143d5453656c6633c0e20a2034190d4b27da3dd137455794704570ec45709139
-size 72093
+oid sha256:d2cec3dbb85679bf7c8dd1a31a9a3a37207d05c1ca1660139102774a858af4d2
+size 55867

data/subsets/mutation_dtm/train.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:675549665c00715760446ed8c08549ffad7a33af06ba9c55a770e6890d5038e1
-size 275341
+oid sha256:fa11942b7d45cedabd64371c7b0a47dd2fff0683d8d10438fce528a48cb74538
+size 282696

data/subsets/mutation_dtm/validation.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3ddd23bb0a3b00f7369b16e91cf3ad8e6f3d75bdfb664cb42239a8e2e1e99428
-size 62706
+oid sha256:b381764db88f1d8f3e73628ef2dbbf3f3dc95feedb85a620ea6480d0f6342a3c
+size 70635

src/05_assign_cluster_splits.py

@@ -45,13 +45,25 @@ def main():
     cl["cluster_id"] = cl["rep"]
     member_to_cluster = cl.set_index("member")["cluster_id"].to_dict()
 
-    # Build protein_id same way as FASTA export
-    df["uniprotkb"] = df["uniprotkb"].astype("string").fillna("").str.strip()
-    df["sequence_id"] = df["sequence_id"].astype("string").fillna("").str.strip()
-
-    df["protein_id"] = df["uniprotkb"]
-    df.loc[df["protein_id"] == "", "protein_id"] = "seqid:" + df.loc[df["protein_id"] == "", "sequence_id"]
-    df.loc[df["protein_id"] == "seqid:", "protein_id"] = "unknown"
+    # Build protein_id robustly
+    for c in ["uniprotkb", "sequence_id", "source_sequence_id", "target_sequence_id", "experiment_id"]:
+        if c not in df.columns:
+            df[c] = pd.NA
+
+    u = df["uniprotkb"].astype("string").fillna("").str.strip()
+    sid = df["sequence_id"].astype("string").fillna("").str.strip()
+    src = df["source_sequence_id"].astype("string").fillna("").str.strip()
+    tgt = df["target_sequence_id"].astype("string").fillna("").str.strip()
+    eid = df["experiment_id"].astype("string").fillna("").str.strip()
+
+    # priority: uniprot > sequence_id > source_sequence_id > target_sequence_id > experiment_id
+    protein_id = u
+    protein_id = protein_id.where(protein_id != "", "seqid:" + sid)
+    protein_id = protein_id.where(protein_id != "seqid:", "srcseq:" + src)
+    protein_id = protein_id.where(protein_id != "srcseq:", "tgtseq:" + tgt)
+    protein_id = protein_id.where(protein_id != "tgtseq:", "exp:" + eid)
+
+    df["protein_id"] = protein_id
 
     df["cluster_id"] = df["protein_id"].map(lambda pid: member_to_cluster.get(pid, None))
 

src/06_make_weighted_splits.py

@@ -0,0 +1,95 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+import argparse
+import pandas as pd
+
+SPLITS = ["train", "validation", "test"]
+
+def assign_weighted_splits(cluster_weights: pd.Series, ratios=(0.8, 0.1, 0.1)) -> pd.DataFrame:
+    total = float(cluster_weights.sum())
+    targets = {
+        "train": total * ratios[0],
+        "validation": total * ratios[1],
+        "test": total * ratios[2],
+    }
+    current = {s: 0.0 for s in SPLITS}
+    assignment = {}
+
+    # Largest-first
+    for cid, w in cluster_weights.sort_values(ascending=False).items():
+        w = float(w)
+
+        # Choose split that minimizes relative fill after adding this cluster.
+        # (current+w)/target; lower is better.
+        def score(s):
+            t = targets[s] if targets[s] > 0 else 1.0
+            return (current[s] + w) / t
+
+        chosen = min(SPLITS, key=score)
+        assignment[cid] = chosen
+        current[chosen] += w
+
+    out = pd.DataFrame({"cluster_id": list(assignment.keys()), "split": list(assignment.values())})
+
+    # Print expected vs achieved (cluster-weighted, i.e. row-weighted)
+    print("Target totals:", {k: round(v) for k, v in targets.items()})
+    print("Achieved totals:", {k: round(v) for k, v in current.items()})
+    return out
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--input", required=True)
+    ap.add_argument("--output", required=True)
+    ap.add_argument("--ratios", default="0.8,0.1,0.1")
+    ap.add_argument("--task", choices=["ddg", "dtm", "binary"], default="ddg")
+    args = ap.parse_args()
+
+    ratios = tuple(float(x) for x in args.ratios.split(","))
+
+    df = pd.read_parquet(args.input)
+    if "cluster_id" not in df.columns:
+        raise ValueError("Input must contain cluster_id")
+
+    # IMPORTANT: remove any existing split so we don't accidentally reuse it
+    if "split" in df.columns:
+        df = df.drop(columns=["split"])
+
+    has_mut = df["mutation"].notna()
+
+    if args.task == "ddg":
+        df_task = df[has_mut & df["ddg"].notna()].copy()
+    elif args.task == "dtm":
+        df_task = df[has_mut & df["dtm"].notna()].copy()
+    else:
+        df_task = df[has_mut & df["stabilizing"].notna()].copy()
+
+    # Ensure cluster_id is a plain string key
+    df_task["cluster_id"] = df_task["cluster_id"].astype("string").fillna("NA_CLUSTER")
+
+    # Cluster weights = number of task rows
+    w = df_task.groupby("cluster_id").size()
+
+    print(f"Task={args.task} rows: {len(df_task):,}")
+    print(f"Task clusters: {len(w):,}")
+    print("Top 10 clusters by rows:")
+    print(w.sort_values(ascending=False).head(10))
+
+    assign = assign_weighted_splits(w, ratios=ratios)
+
+    # Join back to all rows (clusters without task rows -> train by default)
+    df["cluster_id"] = df["cluster_id"].astype("string").fillna("NA_CLUSTER")
+    df = df.merge(assign, on="cluster_id", how="left")
+    df["split"] = df["split"].fillna("train")
+
+    df.to_parquet(args.output, index=False)
+    print("Wrote:", args.output)
+
+    # Quick verify on task rows
+    df_task_out = df_task.merge(assign, on="cluster_id", how="left")
+    df_task_out["split"] = df_task_out["split"].fillna("train")
+    print("Task rows by split:")
+    print(df_task_out["split"].value_counts())
+
+if __name__ == "__main__":
+    main()