segy_to_mdio has an expensive phase where we create a grid. If ingestion fails for any reason (e.g. grid sparsity, _trace_worker failure, etc.) this phase needs to be repeated. Being able to pickle this result would make failure analysis on index headers easier in an MDIO context. It would also help enable resumable ingestions on _trace_worker failure (machine issues).
segy_to_mdiohas an expensive phase where we create a grid. If ingestion fails for any reason (e.g. grid sparsity,_trace_workerfailure, etc.) this phase needs to be repeated. Being able to pickle this result would make failure analysis on index headers easier in an MDIO context. It would also help enable resumable ingestions on_trace_workerfailure (machine issues).