Adapters

PlanFrame is backend-agnostic. Execution is provided by adapters.

Official adapters

• planframe-polars: Polars backend adapter (end-user docs live under the planframe-polars track).
• planframe-pandas: pandas backend adapter (end-user docs live under the planframe-pandas track).
• planframe-sparkless: sparkless backend adapter (end-user docs live under the planframe-sparkless track; provides real execution for the Spark UI).

See also:

• Adapter capability matrix

Adding a new adapter

Start from the core guide:

• Creating an adapter

End-user packages may ship optional API skins (planframe.spark, planframe.pandas) as mixins; adapters only need to implement BaseAdapter (including optional hooks such as hint() if you interpret plan-level hints). See PySpark-like API and pandas-like API.