Module ambit_stochastics.helpers.input_checks

Checks the inputs to the trawl and simple_ambit_field classes.

"""Checks the inputs to the `trawl` and `simple_ambit_field` classes."""

import pandas as pd
import numpy as np

def check1(nr_trawls,nr_simulations):
    #assert (isinstance(tau,float) or isinstance(tau,int)) and tau > 0
    assert isinstance(nr_trawls,int) and nr_trawls >0
    assert isinstance(nr_simulations,int) and nr_simulations >0
    
def check_spatio_temporal_positions(x,tau,k_s,k_t,nr_simulations):
    assert (isinstance(x,float) or isinstance(x,int)) and x > 0
    assert (isinstance(tau,float) or isinstance(tau,int)) and tau > 0
    assert isinstance(k_s,int) and k_s >0
    assert isinstance(k_t,int) and k_t >0
    assert isinstance(nr_simulations,int) and nr_simulations >0


def check_trawl_function(phi):
    """Check if the function is increasing and 0 for strictly positive values"""
    assert callable(phi),'trawl_function is not a function'
    assert phi(0.000001) == 0,'trawl_function does not satisfy trawl_function(t)=0 for t >0'
        
    phi_values = phi(np.linspace(-100,0,10**5))
    assert pd.Series(phi_values).is_monotonic_increasing,'trawl_function is not increasing'   

#distributional checks
        
def check_gaussian_params(gaussian_part_params):
        """Check if the distribution of the jump part is supported and if the parameters 
        of the Gaussian part are numbers"""
        #gaussian part params
        assert isinstance(gaussian_part_params,tuple),'gaussian_part_params is not a tuple'
        assert all(isinstance(i,(int,float)) for i in gaussian_part_params),'parameters of the gaussian part are not numbers'
        
            
def check_jump_part_and_params(jump_part_name,jump_part_params):
        """Check if the distribution of the jump part is supported and if the parameters 
        of the Jump parts are numbers"""
        #jump part params
        assert isinstance(jump_part_params,tuple),'jump_part_params is not a tuple'
        assert all(isinstance(i,(int,float)) for i in jump_part_params),'parameters of the jump part are not numbers'
        
        #jump part_name
        if jump_part_name in ['norminvgauss','geninvgauss','nbinom']: #to also add hyperbolic distributions 
            raise ValueError('distribution not yet supported')

        elif jump_part_name not in [None,'invgauss','gamma','cauchy','poisson']:
            raise ValueError('unknown distribution')
            
def check_grid_params(mesh_size,truncation_grid,times_grid):
    assert  isinstance(mesh_size,(int,float)) and mesh_size >0,'please check mesh size'
    assert  isinstance(truncation_grid,(int,float)) and truncation_grid < 0
            
def check_cpp_params(cpp_part_name, cpp_part_params, cpp_intensity, custom_sampler):
    assert  isinstance(cpp_intensity,(int,float)) and cpp_intensity >0,'please check mesh size' 

    assert cpp_part_name in ['poisson','bernoulli','binom','nbinom','logser','custom'],'cpp_part_name should be one of the \
                following: poisson, binom, nbinom, logser, custom'
    
    if cpp_part_name == 'custom':
        assert callable(custom_sampler),'cpp sampler should be a function'
    
    elif  cpp_part_name in ['poisson','bernoulli','logser']:
        assert isinstance(cpp_part_params,tuple) and len(cpp_part_params) == 1,'cpp_part_params should be a tuple of 1 element' 
        assert isinstance(cpp_part_params[0],(int,float)) and cpp_part_params[0] >=0,'the first element in cpp_part_params should be a non-negative  float'
        
        if cpp_part_name == 'logser':
            assert cpp_part_params[0] < 1, 'cpp_part_params[0] should be strictly less than 1'
        
        elif cpp_part_name == 'bernoulli':
            assert cpp_part_params[0] <=1 , 'cpp_part_params[0] should be less than or equal than 1'
        
    elif cpp_part_name == 'binom' or cpp_part_name == 'nbinom':
        assert isinstance(cpp_part_params,tuple) and len(cpp_part_params) == 2,'cpp_part_params should be a tuple with 2 elements' 
        assert isinstance(cpp_part_params[0],int) and cpp_part_params[0] > 0,'first parameter in cpp_part_params should be a positive integer'
        assert isinstance(cpp_part_params[1],float) and 1 >= cpp_part_params[1] >= 0,'second parameter in cpp_part_params should be a float in [0,1]'
        

    if cpp_part_name != 'custom' and custom_sampler != None:
        raise ValueError('please check whether you are trying to use a custom sampler for \
                         a particular distribution or one of the distributions poisson, binom, nbinom, bernoulli')
        

Functions

def check1(nr_trawls, nr_simulations)

Expand source code

def check1(nr_trawls,nr_simulations):
    #assert (isinstance(tau,float) or isinstance(tau,int)) and tau > 0
    assert isinstance(nr_trawls,int) and nr_trawls >0
    assert isinstance(nr_simulations,int) and nr_simulations >0

def check_cpp_params(cpp_part_name, cpp_part_params, cpp_intensity, custom_sampler)

Expand source code

def check_cpp_params(cpp_part_name, cpp_part_params, cpp_intensity, custom_sampler):
    assert  isinstance(cpp_intensity,(int,float)) and cpp_intensity >0,'please check mesh size' 

    assert cpp_part_name in ['poisson','bernoulli','binom','nbinom','logser','custom'],'cpp_part_name should be one of the \
                following: poisson, binom, nbinom, logser, custom'
    
    if cpp_part_name == 'custom':
        assert callable(custom_sampler),'cpp sampler should be a function'
    
    elif  cpp_part_name in ['poisson','bernoulli','logser']:
        assert isinstance(cpp_part_params,tuple) and len(cpp_part_params) == 1,'cpp_part_params should be a tuple of 1 element' 
        assert isinstance(cpp_part_params[0],(int,float)) and cpp_part_params[0] >=0,'the first element in cpp_part_params should be a non-negative  float'
        
        if cpp_part_name == 'logser':
            assert cpp_part_params[0] < 1, 'cpp_part_params[0] should be strictly less than 1'
        
        elif cpp_part_name == 'bernoulli':
            assert cpp_part_params[0] <=1 , 'cpp_part_params[0] should be less than or equal than 1'
        
    elif cpp_part_name == 'binom' or cpp_part_name == 'nbinom':
        assert isinstance(cpp_part_params,tuple) and len(cpp_part_params) == 2,'cpp_part_params should be a tuple with 2 elements' 
        assert isinstance(cpp_part_params[0],int) and cpp_part_params[0] > 0,'first parameter in cpp_part_params should be a positive integer'
        assert isinstance(cpp_part_params[1],float) and 1 >= cpp_part_params[1] >= 0,'second parameter in cpp_part_params should be a float in [0,1]'
        

    if cpp_part_name != 'custom' and custom_sampler != None:
        raise ValueError('please check whether you are trying to use a custom sampler for \
                         a particular distribution or one of the distributions poisson, binom, nbinom, bernoulli')

def check_gaussian_params(gaussian_part_params)

Check if the distribution of the jump part is supported and if the parameters of the Gaussian part are numbers

Expand source code

def check_gaussian_params(gaussian_part_params):
        """Check if the distribution of the jump part is supported and if the parameters 
        of the Gaussian part are numbers"""
        #gaussian part params
        assert isinstance(gaussian_part_params,tuple),'gaussian_part_params is not a tuple'
        assert all(isinstance(i,(int,float)) for i in gaussian_part_params),'parameters of the gaussian part are not numbers'

def check_grid_params(mesh_size, truncation_grid, times_grid)

Expand source code

def check_grid_params(mesh_size,truncation_grid,times_grid):
    assert  isinstance(mesh_size,(int,float)) and mesh_size >0,'please check mesh size'
    assert  isinstance(truncation_grid,(int,float)) and truncation_grid < 0

def check_jump_part_and_params(jump_part_name, jump_part_params)

Check if the distribution of the jump part is supported and if the parameters of the Jump parts are numbers

Expand source code

def check_jump_part_and_params(jump_part_name,jump_part_params):
        """Check if the distribution of the jump part is supported and if the parameters 
        of the Jump parts are numbers"""
        #jump part params
        assert isinstance(jump_part_params,tuple),'jump_part_params is not a tuple'
        assert all(isinstance(i,(int,float)) for i in jump_part_params),'parameters of the jump part are not numbers'
        
        #jump part_name
        if jump_part_name in ['norminvgauss','geninvgauss','nbinom']: #to also add hyperbolic distributions 
            raise ValueError('distribution not yet supported')

        elif jump_part_name not in [None,'invgauss','gamma','cauchy','poisson']:
            raise ValueError('unknown distribution')

def check_spatio_temporal_positions(x, tau, k_s, k_t, nr_simulations)

Expand source code

def check_spatio_temporal_positions(x,tau,k_s,k_t,nr_simulations):
    assert (isinstance(x,float) or isinstance(x,int)) and x > 0
    assert (isinstance(tau,float) or isinstance(tau,int)) and tau > 0
    assert isinstance(k_s,int) and k_s >0
    assert isinstance(k_t,int) and k_t >0
    assert isinstance(nr_simulations,int) and nr_simulations >0

def check_trawl_function(phi)

Check if the function is increasing and 0 for strictly positive values

Expand source code

def check_trawl_function(phi):
    """Check if the function is increasing and 0 for strictly positive values"""
    assert callable(phi),'trawl_function is not a function'
    assert phi(0.000001) == 0,'trawl_function does not satisfy trawl_function(t)=0 for t >0'
        
    phi_values = phi(np.linspace(-100,0,10**5))
    assert pd.Series(phi_values).is_monotonic_increasing,'trawl_function is not increasing'