Asset Loading

load_rocky_assets() calls rocky discover and returns a list of Dagster AssetSpec objects, one for each enabled table across all sources. This gives you automatic asset discovery without manually defining each table.

How it works

1. load_rocky_assets() invokes rocky.discover() under the hood.
2. Each enabled table across all sources becomes one AssetSpec.
3. Asset keys, groups, tags, and metadata are derived from the source and table information.

Default mappings

Asset key

The default asset key is constructed from the source type, component values, and table name:

[source_type, *component_values, table_name]

For example, a table orders from a Fivetran source with components tenant=acme, regions=us_west, connector=shopify produces:

["fivetran", "acme", "us_west", "shopify", "orders"]

Group

The default group name is the first component value. For the example above, the group would be "acme".

Tags

• rocky/source_type — the source type (e.g., "fivetran")
• rocky/<component_name> — one tag per string component (e.g., rocky/tenant: "acme")

Metadata

• source_id — the source identifier
• source_type — the source type
• last_sync_at — timestamp of the last sync
• row_count — number of rows in the table

Example

from dagster_rocky import RockyResource, load_rocky_assets
import dagster as dg

rocky = RockyResource(config_path="rocky.toml")
assets = load_rocky_assets(rocky)

defs = dg.Definitions(
    assets=assets,
    resources={"rocky": rocky},
)

Custom translation

To customize how sources and tables map to Dagster concepts, pass a custom translator. See Translator for details.

assets = load_rocky_assets(rocky, translator=MyTranslator())