Data Readiness Workflow#

This is a V0.29 recipe page over existing public APIs; it adds no runtime surface.

Use this path when you have PDE time-series data and want to know whether it can enter the stable PDELie stack.

xarray.Dataset or array-like data
-> readiness report
-> explicit scalar FieldBatch conversion
-> optional residual preflight

Recipe#

  1. Inspect the source container before conversion.

  2. Select one scalar data variable explicitly when the input is an xarray.Dataset.

  3. Provide canonical metadata yourself; Dataset attrs are report metadata, not trusted PDE identity.

  4. Convert to FieldBatch only after the readiness report is acceptable.

  5. Run summarize_field_batch_readiness(...) with the residual evaluator you intend to use.

Primary APIs#

  • pdelie.reporting.summarize_xarray_dataset_readiness(...)

  • pdelie.data.from_xarray_dataset(...)

  • pdelie.data.from_numpy(...)

  • pdelie.data.from_xarray(...)

  • pdelie.reporting.summarize_field_batch_readiness(...)

Defensible Claim#

A ready report means the configured checks passed for the current scalar 1D uniform periodic contract. It does not infer PDE identity, resample nonuniform grids, load files, or make broad external-data claims.

Next Step#

Continue to end_to_end_dataset_to_downstream.md for a full recipe that carries a Dataset through residual preflight, generator confidence, and downstream discovery summaries.