Get a data lake file

Returns a lake member's ingestion status. Adding a file to a lake does NOT ingest it: a file only becomes searchable once it has been chunked and embedded, so poll this endpoint until ingestion_status is ready. indexing clears on its own; paused means re-processing was stopped partway and needs the file reprocessed (or background lake work resumed) before it returns; failed carries the reason in error. A ready file can also carry a non-null error when a later pass failed partway: its committed passages stay searchable, so branch on ingestion_status, not on error. not_ingested means the file has no searchable passages yet: ingestion has not started, a first pass has not committed its chunks, or the file has no extractable text (such a file never becomes ready). These agree with what search reports under retrieval_unavailable.