Utilities

def regressor_coefficients(m):
    """
    Summarize regressor coefficients with confidence intervals.

    Parameters:
    - m: Prophet, fitted Prophet model with regressors

    Returns:
    - DataFrame with regressor names, coefficients, and confidence intervals
    """

def regressor_index(m, name):
    """
    Get the column index of a regressor in the beta matrix.

    Parameters:
    - m: Prophet, fitted Prophet model
    - name: str, name of the regressor

    Returns:
    - int, column index of the regressor in model parameters
    """

def warm_start_params(m):
    """
    Retrieve parameters from a fitted model for warm-starting a new model.

    Parameters:
    - m: Prophet, fitted Prophet model

    Returns:
    - dict, dictionary of parameters that can be used to initialize a new model
    """

from prophet import Prophet
from prophet.utilities import regressor_coefficients, regressor_index

# Create model with regressors
model = Prophet()
model.add_regressor('temperature')
model.add_regressor('humidity')

# Fit model with regressor data
model.fit(df_with_regressors)

# Analyze regressor coefficients
coeffs = regressor_coefficients(model)
print(coeffs)
# Output shows coefficient estimates and confidence intervals

# Get specific regressor index
temp_idx = regressor_index(model, 'temperature')
print(f"Temperature regressor is at index: {temp_idx}")

from prophet.utilities import warm_start_params

# Train initial model
model1 = Prophet()
model1.fit(training_data)

# Extract parameters for warm start
params = warm_start_params(model1)

# Create new model with similar configuration
model2 = Prophet(
    changepoint_prior_scale=model1.changepoint_prior_scale,
    seasonality_prior_scale=model1.seasonality_prior_scale
)

# The warm start parameters can be used to initialize
# the new model's internal state for faster fitting
model2.fit(extended_training_data)

# Examine all regressor effects
if model.extra_regressors:
    print("Model has external regressors:")
    for name, regressor_info in model.extra_regressors.items():
        idx = regressor_index(model, name)
        print(f"  {name}: index {idx}, mode {regressor_info['mode']}")
    
    # Get coefficient summary
    coeff_summary = regressor_coefficients(model)
    print("\nRegressor Coefficients:")
    print(coeff_summary)
else:
    print("Model has no external regressors")