OptionalbytesSize of the stored file dataUrl points at — a converted lastra for a CSV/gzip/zip upload, or the parquet file itself, unconverted, for a parquet upload. Not the size of the bytes originally PUT to storage; see dataFormat.
OptionalcadenceThe cadence discovered from the data's own timestamps. Either a fixed grid — 1s,
5s, 15s, 1m, 5m, 15m, 30m, 1h, 4h, 1d — when at least half the
intervals between consecutive rows fall on that step (small clock jitter tolerated), or
rt: native data at the rate it was captured, each row at its own timestamp with no
fixed step — per-trade on-chain swaps, block-spaced or sub-second ticks, irregular
intervals. An rt dataset can be resampled to any fixed cadence at prepare time.
OptionaldataWhich format dataUrl is actually in — check this rather than assuming it matches
how you uploaded it. lastra — our native columnar format — for a CSV (or gzip/zip
of one) upload, always converted on ingest, or for a lastra upload, stored as-is
(the value alone doesn't tell you which). parquet for a parquet upload, also
stored as-is today.
OptionaldataPresigned GET URL to the stored file — see dataFormat for which format it's in.
Present once the version is ready. Long-lived (day-scale, not permanent): a
DuckDB-WASM/lastra-ts-style reader issues HTTP range requests against it lazily over
an extended viewing session, not in one shot like a browser upload.
OptionalgapsNumber of gaps at the discovered cadence. Always 0 for rt.
OptionalidThe version id. Pass as datasetVersionId on POST .../prepare to pin it.
OptionallargestThe largest gap, in units of the discovered cadence step. Always 0 for rt.
OptionalrowsNumber of data rows.
OptionaltimestampThe unit the timestamp column was uploaded in — ISO-8601, or the epoch band its
numeric values fell in (seconds, millis, or micros).
One successfully ingested upload. Cadence and timestamp unit are discovered from the file, not declared by the caller.