Generation

The generation API returns typed in-memory objects by default and can persist the same deterministic run when write=True.

generate(..., write=False) returns fraudtwin.GeneratedData, whose entities and behavior attributes are available for graph and dataset workflows. generate(..., write=True) returns fraudtwin.GeneratedRun, a lightweight manifest/path object suitable for large runs. Call run.load_data() when a persisted run must be loaded back into typed domain records. If a point-in-time dataset is enabled, use data.require_dataset() to narrow the optional dataset and enable IDE completion for rows, count, and frame.

fraudtwin.generate	Generate a deterministic FraudTwin run.
fraudtwin.generate_scale	Run an explicitly requested scale job using the scale manifest path.
fraudtwin.iter_scale_run	Yield scale records from a producer or an existing partitioned run.
fraudtwin.iter_scale_records	Stream canonical entities, profiles, payments, events, and ledger rows.
fraudtwin.resume_generation	Resume a scale run from a validated checkpoint manifest.
fraudtwin.GeneratedData	Generated FraudTwin data kept in memory.
fraudtwin.GeneratedRun	Metadata and paths for a generated run written to disk.

Large-scale iteration

fraudtwin.generate_scale	Run an explicitly requested scale job using the scale manifest path.
fraudtwin.iter_scale_run	Yield scale records from a producer or an existing partitioned run.
fraudtwin.iter_scale_records	Stream canonical entities, profiles, payments, events, and ledger rows.
fraudtwin.run_scale_benchmark	Run one explicitly requested scale job and publish machine evidence.
fraudtwin.write_scale_benchmark_manifest	Write non-deterministic machine evidence separately from run identity.