quantlab.acquisition.registry.refresh_raw_tier#

quantlab.acquisition.registry.refresh_raw_tier(descriptor, download_dir, *, end=None, **options)[source]#

Keep a vendor’s raw tier current from its watermarks, in the current process.

Calls the refresh the vendor registered on its descriptor (SourceDescriptor.raw_tier_refresh, a quantlab.acquisition.refresh.RawTierRefresh), so the caller names a vendor, never a client: Sharadar’s daily table plan, Massive’s REST endpoints, FRED’s series. It writes the raw tier only; the stores are brought up to date afterwards from it.

This is the registry’s one way to keep a raw tier current (ADR 0049). A vendor whose download is symbol-batched serves it by wrapping its SymbolBatchAcquisition in its refresh, as FRED does; downloading an explicit roster over an explicit window is download_symbols().

Parameters:

descriptorSourceDescriptor: The vendor to refresh.
download_dirstr or Path: The download root; the vendor’s raw tier sits under it.
enddate or None, default None: The last day to pull; None is the vendor’s today.
**options: The vendor’s own settings, passed through (Sharadar’s trading_days; Massive’s endpoints; FRED’s series and start).

Returns:

RawTierRefreshResult: Every part tried, a failed one with its error.

Raises:

ValueError: If the vendor registered no refresh (Tiingo, Alpaca, WRDS).

Examples

Needs the vendor’s credential in the environment and network access:

from quantlab.acquisition.base import DataSourceRegistry
from quantlab.acquisition.registry import refresh_raw_tier

result = refresh_raw_tier(DataSourceRegistry.get("massive"), "/data/quantlab/downloads")
result.failed  # the endpoints not refreshed

A vendor without one is refused before anything is touched:

>>> from quantlab.acquisition.base import DataSourceRegistry
>>> refresh_raw_tier(DataSourceRegistry.get("tiingo"), "downloads")
Traceback (most recent call last):
...
ValueError: Tiingo EOD registers no raw-tier refresh; download it with download_symbols().