Remove UTF-8 BOM from file stream

### Summary of the new feature / enhancement

When reading a file, it might have the UTF-8 BOM  (0xEF, 0xBB, 0xBF).  This currently causes deserialization to fail.

### Proposed technical implementation details (optional)

Before deserialization, we can probably check if the first 3 bytes is the BOM and remove it.