Hi team, could someone please review PR #7576? It adds

daft.read_orc()

for local and remote ORC files, with lazy batched reads, schema alignment, and support for paths, directories, and globs. This lets users process Hive/Spark ORC data directly in Daft without a separate read-and-materialize or format-conversion step. Feedback on the API and edge cases would be very welcome. Thanks!

Hey hey, I'm taking a look now 🙂