fix: convert memoryview to bytes for PostgreSQL blob unpacking

psycopg2 returns bytea columns as memoryview objects, which lack the
startswith() method needed by the blob decompression code. Convert to
bytes at the start of unpack() for compatibility.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: dimitri-yatsenko

src/datajoint/blob.py

@@ -149,6 +149,9 @@ def squeeze(self, array: np.ndarray, convert_to_scalar: bool = True) -> np.ndarr
         return array.item() if array.ndim == 0 and convert_to_scalar else array
 
     def unpack(self, blob):
+        # PostgreSQL returns bytea as memoryview; convert to bytes for string operations
+        if isinstance(blob, memoryview):
+            blob = bytes(blob)
         self._blob = blob
         try:
             # decompress