Upload 3 files

anonymize_mlflow.py

@@ -0,0 +1,53 @@
+#!/usr/bin/env python3
+import os
+import re
+import glob
+
+def anonymize_file(file_path):
+    """Anonymize personal data in files."""
+    with open(file_path, 'r') as f:
+        lines = f.readlines()
+    
+    modified = False
+    for i, line in enumerate(lines):
+        # Anonymize artifact_uri or artifact_location
+        if 'artifact_uri:' in line or 'artifact_location:' in line:
+            new_line = re.sub(r'file://(.+?)(/mlruns/)', r'file://<placeholder>\2', line)
+            if new_line != line:
+                lines[i] = new_line
+                modified = True
+        # Anonymize user_id
+        elif 'user_id:' in line:
+            new_line = re.sub(r'user_id:\s*.+', 'user_id: <user_placeholder>', line)
+            if new_line != line:
+                lines[i] = new_line
+                modified = True
+        # Anonymize data_dir and csv_path in config files
+        elif 'data_dir:' in line or 'csv_path:' in line:
+            new_line = re.sub(r':\s*(.+?)/pdmx/', r': pdmx/', line)
+            if new_line != line:
+                lines[i] = new_line
+                modified = True
+    
+    if modified:
+        with open(file_path, 'w') as f:
+            f.writelines(lines)
+        print(f"Anonymized: {file_path}")
+
+def main():
+    # Find all relevant files recursively
+    yaml_files = glob.glob('**/*.yaml', recursive=True)
+    for file_path in yaml_files:
+        anonymize_file(file_path)
+    
+    # Also anonymize mlflow.user tag files
+    user_files = glob.glob('**/tags/mlflow.user', recursive=True)
+    for file_path in user_files:
+        with open(file_path, 'w') as f:
+            f.write('<user_placeholder>\n')
+        print(f"Anonymized: {file_path}")
+    
+    print("Anonymization complete.")
+
+if __name__ == "__main__":
+    main()

mlruns.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b7e943d91f708ea6291d208a3f97705c7851961fa3efecb456d95637f07e511
+size 1558271080

restore_mlflow_paths.py

@@ -0,0 +1,58 @@
+import re
+import glob
+import sys
+
+def restore_file(file_path, absolute_path, user_id):
+    """Restore placeholders in files with the specified absolute path and user ID."""
+    with open(file_path, 'r') as f:
+        lines = f.readlines()
+    
+    modified = False
+    for i, line in enumerate(lines):
+        # Restore artifact_uri or artifact_location
+        if 'artifact_uri:' in line or 'artifact_location:' in line:
+            new_line = re.sub(r'file://<placeholder>(/mlruns/)', rf'file://{absolute_path}\1', line)
+            if new_line != line:
+                lines[i] = new_line
+                modified = True
+        # Restore user_id
+        elif 'user_id:' in line and '<user_placeholder>' in line:
+            new_line = re.sub(r'user_id:\s*<user_placeholder>', f'user_id: {user_id}', line)
+            if new_line != line:
+                lines[i] = new_line
+                modified = True
+    
+    if modified:
+        with open(file_path, 'w') as f:
+            f.writelines(lines)
+        print(f"Restored: {file_path}")
+
+def main():
+    if len(sys.argv) != 3:
+        print("Usage: python restore_mlflow_paths.py <absolute_path> <user_id>")
+        print("Example: python restore_mlflow_paths.py /home/user/path/to/mlruns hendrik-roth")
+        sys.exit(1)
+    
+    absolute_path = sys.argv[1]
+    user_id = sys.argv[2]
+    # Ensure the path starts with / and does not end with /
+    if not absolute_path.startswith('/'):
+        absolute_path = '/' + absolute_path
+    absolute_path = absolute_path.rstrip('/')
+    
+    # Find all yaml files recursively
+    yaml_files = glob.glob('**/*.yaml', recursive=True)
+    for file_path in yaml_files:
+        restore_file(file_path, absolute_path, user_id)
+    
+    # Also restore mlflow.user tag files
+    user_files = glob.glob('**/tags/mlflow.user', recursive=True)
+    for file_path in user_files:
+        with open(file_path, 'w') as f:
+            f.write(f'{user_id}\n')
+        print(f"Restored: {file_path}")
+    
+    print("Path restoration complete.")
+
+if __name__ == "__main__":
+    main()