Skip to main content

Versioning & Changelogs

How Snowtrail versions data and tracks changes.

API Versioning​

The Snowtrail API is accessed at:

https://api.snowtrail.ai/{product_id}/{dataset_name}

There is no version prefix in the URL. When breaking changes are introduced, they will be communicated in advance via the changelog and migration guides. (Versioning will be introduced in the near future).

Data Versioning​

Every derived row carries a model_version column recording the version of the model that produced it, in MAJOR.MINOR form:

  • MAJOR increments when a change alters historical values, a redefined feature, a recalibrated threshold, a corrected computation. Values you retrieved previously may now differ.
  • MINOR increments for additive or non-value-affecting changes, a new diagnostic column, expanded coverage, documentation, performance work.

Comparing model_version between two pulls tells you whether a series is still comparable with one you already hold.

Because our derived layers are deterministic recomputations of raw data, a MAJOR version change restates history: the table is rebuilt end-to-end under the new definition. Rows from a superseded MAJOR are outranked by the new MAJOR in every query, including as_of, and may be retired. model_version tells you that a restatement occurred and what produced the current values; it does not provide access to the superseded ones.

from snowtrail import Snowtrail

client = Snowtrail(api_key="your-api-key")

# Access data using product-specific methods
df = client.gbsi_us.system_stress(date_from="2024-01-01")