Upload data1/rename.py with huggingface_hub

data1/rename.py

@@ -0,0 +1,52 @@
+import os
+import pandas as pd
+from pathlib import Path
+
+
+def rename_repos(csv_path: str, repos_dir: str):
+    """Match and rename repos from various formats to owner___repo"""
+    repos_dir = Path(repos_dir)
+    df = pd.read_csv(csv_path)
+
+    # Build lookup: possible old names -> new name (owner___repo)
+    lookup = {}
+    for _, row in df.iterrows():
+        owner, repo = row["full_name"].split("/")
+        new_name = f"{owner}___{repo}"  # three underscores
+        # Possible old patterns
+        lookup[repo] = new_name  # just repo name
+        lookup[f"{owner}_{repo}"] = new_name  # owner_repo (single underscore)
+        lookup[f"{owner}__{repo}"] = new_name  # owner__repo (double underscore, just in case)
+        lookup[new_name] = new_name  # already correct
+
+    # Get existing directories
+    existing = [d for d in repos_dir.iterdir() if d.is_dir() and not d.name.startswith(".")]
+
+    renamed, skipped = 0, 0
+    for d in existing:
+        if d.name in lookup:
+            new_name = lookup[d.name]
+            if d.name != new_name:
+                new_path = repos_dir / new_name
+                if new_path.exists():
+                    print(f"[SKIP] Target exists: {d.name} -> {new_name}")
+                    skipped += 1
+                else:
+                    d.rename(new_path)
+                    print(f"[OK] {d.name} -> {new_name}")
+                    renamed += 1
+        else:
+            print(f"[WARN] Not in CSV: {d.name}")
+            skipped += 1
+
+    print(f"\nDone: {renamed} renamed, {skipped} skipped")
+
+
+if __name__ == "__main__":
+    import argparse
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--csv", default="./workdir/repos_checked.csv")
+    parser.add_argument("--repos", default="./workdir/repos_raw")
+    args = parser.parse_args()
+    rename_repos(args.csv, args.repos)