Selecting a class method based on a random number

Numba Community Support
1

Hi everyone,

I have a simple object-oriented python + njit program, which also has a funny error - it is a TypingError, but I have no idea how to solve it!

Detail: I want to index a list based on a number i, 0 or 1. The choice is based on a condition (a function); if I used the return i to index to list (R[i]), it does not work! But if I say: if i == 0: R[0], works!

Error triggered:
numba.core.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
Cannot getitem on a literal list, return type cannot be statically determined.

If someone can figure out why…! See the code below!

import numpy as np
import matplotlib.pyplot as plt
import time
from   numba import jit
from   numba.experimental import jitclass
from   numba import types, typed, int64   
from   numba.types import ListType
from   numba.typed.typedlist import List
import random

spec = [('detail', int64),]
@jitclass(spec)
class Species:
    def __init__(self, detail = 0):
        self.detail = int(detail)

@jitclass
class Actions:
    def __init__(self):
        pass

    def OptionA(self, A, B):
        A.detail -= 1
        B.detail += 2
        return A, B
    
    def OptionB(self, A, B, C):
        A.detail  -= 1   
        B.detail  -= 1    
        C.detail  += 1    
        return A, B, C

@jit(nopython=True)
def cumsum(x, target):    
    total  = 0
    for i, x_i in enumerate(x):
        total += x_i
        if total >= target:
            return i
    raise IndexError('list index out of range')

@jit(nopython=True)
def TestNumba(species, s_actions):

    SpecieA, SpecieB, SpecieC   = species

    r    = [[s_actions.OptionA, (SpecieA, SpecieB)],
            [s_actions.OptionB, (SpecieA, SpecieB, SpecieC)]]
    
    actions_options = np.array([SpecieA.detail*0.5, SpecieA.detail*SpecieB.detail*2.3])

    t    = 0 
    tend = 10 

    while t < tend:
        actions_options_sum  = np.sum(actions_options) 
        rnd                  = np.random.uniform(0.0, 1.0)*actions_options_sum
        i                    = cumsum(actions_options, rnd)

        r[i][0](*r[i][1])

        ## Like this works:
        # if i == 0:
        #     r[0][0](*r[0][1])
        # elif i == 1:
        #     r[1][0](*r[1][1])

        t += 1
    
    print(SpecieA.detail, SpecieB.detail, SpecieC.detail)
    return


SpecieA = Species(50)
SpecieB = Species(200)
SpecieC = Species()

species         = [SpecieA, SpecieB, SpecieC]
s_actions       = Actions()

startime = time.time()              
TestNumba(species, s_actions)
endtime  = time.time() - startime    ```
2

I have also been getting the same error if I’m selecting functions: I keep getting. a typing error

No implementation of function Function(< built-in function getitem >) found