API reference¶

The reference below is generated on every build from the Python source and its NumPy-style docstrings. Signatures are never maintained by hand. Zensical's supported API navigation discovers the package; mkdocstrings renders its members. Source changes are also watched during local preview.

All ten features are available from curvefit_stats and directly from their functional modules. There are exactly two functional submodules:

• Inference: bands, diagnostics, parameter tables, profile likelihood, bootstrap, collinearity, conditional covariance and propagation.
• Estimators: shared-parameter global fitting and EIV.

Read the statistical contract¶

A covariance array does not retain its noise model, estimator, residual degrees of freedom or active constraints. Keep that context when combining APIs. Start with the workflow guide and cross-module integration.

Type aliases denote float64 NumPy arrays or callables accepting func(x, *parameters). Result dataclass names beginning with an underscore are implementation details; access the documented result attributes rather than importing private classes. Exceptions raised by your own model or derived callables can propagate in addition to each API's listed errors.