Implement repr() for ArrayRecordDataSource.

PiperOrigin-RevId: 516235396

python/array_record_data_source.py

@@ -34,6 +34,7 @@ def __getitem__(self, record_keys: Sequence[int]) -> Sequence[T]:
 import bisect
 from concurrent import futures
 import dataclasses
+import hashlib
 import itertools
 import os
 import pathlib
@@ -310,6 +311,13 @@ def close(self):
         reader.close()
         self._readers[reader_idx] = None
 
+  def __repr__(self) -> str:
+    """Storing a hash of paths since paths can be a very long list."""
+    h = hashlib.sha1()
+    for p in self._paths:
+      h.update(p.encode())
+    return f"ArrayRecordDataSource(hash_of_paths={h.hexdigest()})"
+
 
 def _get_flag_value(flag: flags.FlagHolder[int]) -> int:
   """Retrieves the flag value or the default if run outside of absl."""