API Reference

Complete API reference for PhyTorch modules and functions.

Unified API

PhyTorch provides a simple, consistent interface for all models:

from phytorch import fit

result = fit(model, data, options=None)

Parameters

• model: Any PhyTorch model instance
• data: Dictionary mapping variable names to numpy arrays
• options: Optional FitOptions object for customization

Returns

• FitResult: Object containing fitted parameters, predictions, statistics, and plotting methods

Core Function

fit(model, data, options=None)

Universal fitting function for all PhyTorch models.

from phytorch import fit
from phytorch.models.generic import Sigmoidal

data = {'x': x_data, 'y': y_data}
result = fit(Sigmoidal(), data)

# Access results
print(result.parameters)  # Fitted parameter values
print(result.r_squared)   # Goodness of fit
predictions = result.predict(new_data)
result.plot()  # Automatic visualization

Available Models

Generic Models (phytorch.models.generic)

General-purpose curve fitting models:

Model	Description	Required Data
Linear	Linear regression y = a*x + b	x, y
Sigmoidal	Rational sigmoid curve	x, y
RectangularHyperbola	Michaelis-Menten kinetics	x, y
NonrectangularHyperbola	Non-rectangular hyperbola	x, y
Arrhenius	Temperature response	x, y
PeakedArrhenius	Peaked temperature response	x, y
Gaussian	Bell-shaped curve	x, y
Weibull	Weibull distribution PDF	x, y
Beta	Beta distribution	x, y

Example:

from phytorch import fit
from phytorch.models.generic import RectangularHyperbola

data = {'x': substrate_conc, 'y': reaction_rate}
result = fit(RectangularHyperbola(), data)

Hydraulics Models (phytorch.models.hydraulics)

Plant water relations models:

Model	Description	Required Data
Sigmoidal	Hydraulic vulnerability curve	x, psi
SJB2018	Pressure-volume curve (Sack, John, Buckley 2018)	w, psi

Example:

from phytorch import fit
from phytorch.models.hydraulics import SJB2018

data = {'w': relative_water_content, 'psi': water_potential}
result = fit(SJB2018(), data)
print(f"Turgor loss point: {result.parameters['w_tlp']:.3f}")

Photosynthesis Models (phytorch.models.photosynthesis)

Leaf gas exchange models:

Model	Description	Required Data
FvCB	Farquhar-von Caemmerer-Berry C3 photosynthesis	Ci, Q, Tleaf/T, A

Example:

from phytorch import fit
from phytorch.models.photosynthesis import FvCB

data = {
    'Ci': intercellular_co2,
    'Q': light_intensity,
    'Tleaf': leaf_temperature,
    'A': net_photosynthesis
}
result = fit(FvCB(), data)
result.plot()  # Generates comprehensive photosynthesis plots

FitResult Object

The FitResult object returned by fit() contains:

Attributes

• parameters: Dictionary of fitted parameter values
• r_squared: Coefficient of determination
• rmse: Root mean squared error
• data: Original data used for fitting
• model: The fitted model instance

Methods

• predict(data): Generate predictions for new data
• plot(save=None, show=True): Visualize fit results

# Using FitResult
result = fit(model, data)

# Access fitted parameters
vcmax = result.parameters['Vcmax25']

# Make predictions
predictions = result.predict(new_data)

# Plot results
result.plot(save='my_fit.png')

FitOptions

Customize the fitting process:

from phytorch import fit, FitOptions

options = FitOptions(
    optimizer='scipy',           # 'scipy' or 'adam'
    maxiter=5000,                # Maximum iterations
    bounds={'param': (0, 100)}   # Parameter bounds
)

result = fit(model, data, options)

FitOptions Parameters

• optimizer: Optimizer to use ('scipy', 'adam')
• maxiter: Maximum number of iterations
• bounds: Dictionary of parameter bounds {param: (lower, upper)}
• initial_guess: Dictionary of initial parameter values

Model Base Class

All models inherit from the Model base class and implement:

class Model:
    def forward(self, data: dict, parameters: dict) -> np.ndarray:
        """Compute model predictions"""
        pass

    def parameter_info(self) -> dict:
        """Return parameter bounds and metadata"""
        pass

    def required_data(self) -> list:
        """Return list of required data fields"""
        pass

    def initial_guess(self, data: dict) -> dict:
        """Generate initial parameter estimates"""
        pass

Quick Reference

Fitting a Model

from phytorch import fit
from phytorch.models.generic import Sigmoidal

result = fit(Sigmoidal(), {'x': x_data, 'y': y_data})

Plotting Results

result.plot()                    # Show plot
result.plot(save='fit.png')      # Save plot
result.plot(save='fit.png', show=False)  # Save without showing

Making Predictions

new_data = {'x': new_x_values, 'y': new_y_values}
predictions = result.predict(new_data)