src/split_dataset.py

import pandas as pd
import shutil
import os
import glob

# ============================================================
# CONFIGURATION - Edit these paths as needed
# ============================================================
PDB_ROOT = "pdb"
OUTPUT_ROOT = "pdb_2021aug02"

TRAIN_CSV      = "train.csv"
VALIDATION_CSV = "validation.csv"
TEST_CSV       = "test.csv"

SPLITS = {
    "training":   TRAIN_CSV,
    "validation": VALIDATION_CSV,
    "test":       TEST_CSV
}

# ============================================================
# HELPER FUNCTIONS
# ============================================================

def get_subfolder(name):
    """
    Extract subfolder from filename.
    E.g., '1nb4_A' -> 'nb'  (characters at index 1 and 2)
    """
    # Strip extension if present
    base = os.path.splitext(name)[0]
    return base[1:3].lower()  # e.g., '1nb4_A' -> 'nb'



def read_csv_entries(csv_file):
    """
    Read comma-separated entries from a single-line CSV.
    E.g., 5naf_A,5naf_B,5naf_C,...
    """
    entries = []
    with open(csv_file, 'r') as f:
        for line in f:
            entries.extend(line.strip().split(','))
    return [e.strip() for e in entries if e.strip()]

def find_pt_file(pdb_root, name):
    """
    Locate the .pt file given the entry name.
    E.g., '1nb4_A' -> pdb_root/nb/1nb4_A.pt
    """
    subfolder = get_subfolder(name)
    pt_path = os.path.join(pdb_root, subfolder, f"{name}.pt")
    if os.path.exists(pt_path):
        return pt_path, subfolder
    return None, subfolder


def copy_file(src, dst_dir, filename):
    """
    Copy file to destination directory, creating dirs if needed.
    """
    os.makedirs(dst_dir, exist_ok=True)
    dst = os.path.join(dst_dir, filename)
    shutil.copy2(src, dst)
    return dst


# ============================================================
# MAIN
# ============================================================

def main():
    stats = {}

    for split_name, csv_file in SPLITS.items():
        print(f"\n{'='*50}")
        print(f"Processing: {split_name.upper()} <- {csv_file}")
        print(f"{'='*50}")

        if not os.path.exists(csv_file):
            print(f"  [WARNING] CSV file not found: {csv_file}")
            continue

        entries = read_csv_entries(csv_file)
        print(f"  Total entries found in CSV: {len(entries)}")

        copied  = 0
        missing = []

        for name in entries:
            src_path, subfolder = find_pt_file(PDB_ROOT, name)

            if src_path:
                # Destination: e.g., OUTPUT_ROOT/validation/nb/1nb4_A.pt
                dst_dir = os.path.join(OUTPUT_ROOT, split_name, subfolder)
                copy_file(src_path, dst_dir, f"{name}.pt")
                copied += 1
            else:
                missing.append(name)

        print(f"Copied  : {copied}")
        print(f"Missing : {len(missing)}")

        if missing:
            missing_log = os.path.join(OUTPUT_ROOT, f"{split_name}_missing.txt")
            with open(missing_log, "w") as f:
                f.write("\n".join(missing))
            print(f"Missing entries logged to: {missing_log}")

        stats[split_name] = {"copied": copied, "missing": len(missing)}

    # ---- Summary ----
    print(f"\n{'='*50}")
    print("SUMMARY")
    print(f"{'='*50}")
    for split, s in stats.items():
        print(f"  {split:<12}: {s['copied']} copied, {s['missing']} missing")


if __name__ == "__main__":
    main()