Wrong number or type of arguments for overloaded function 'VariableElimination_addEvidence'

[kenneth-lee-ch]

@phwuil Hi,

I occasionally get the following error when I call the functiongum.getPosterior(bn, evs= evidence, target=self.target) , do you know what that means?

File "/Users/lee4094/miniconda3/envs/invariant-prediction/lib/python3.9/site-packages/pyAgrum/__init__.py", line 556, in getPosterior
    inf.setEvidence(evs)
  File "/Users/lee4094/miniconda3/envs/invariant-prediction/lib/python3.9/site-packages/pyAgrum/pyAgrum.py", line 15296, in setEvidence
    self.addEvidence(k,v)
  File "/Users/lee4094/miniconda3/envs/invariant-prediction/lib/python3.9/site-packages/pyAgrum/pyAgrum.py", line 15607, in addEvidence
    return _pyAgrum.VariableElimination_addEvidence(self, *args)
TypeError: in method 'VariableElimination_addEvidence', argument 3 of type 'std::vector< double,std::allocator< double > > const &'
Additional information:
Wrong number or type of arguments for overloaded function 'VariableElimination_addEvidence'.
  Possible C/C++ prototypes are:
    gum::VariableElimination< double >::addEvidence(gum::NodeId const,gum::Idx const)
    gum::VariableElimination< double >::addEvidence(std::string const &,gum::Idx const)
    gum::VariableElimination< double >::addEvidence(gum::NodeId const,std::string const &)
    gum::VariableElimination< double >::addEvidence(std::string const &,std::string const &)
    gum::VariableElimination< double >::addEvidence(gum::NodeId const,std::vector< double,std::allocator< double > > const &)
    gum::VariableElimination< double >::addEvidence(std::string const &,std::vector< double,std::allocator< double > > const &)

[phwuil]

Hi Kenneth, can you show the code that causes this error ?

[kenneth-lee-ch]

@phwuil Here is the bif file of the bn: https://drive.google.com/file/d/1I-3migzH5ZwIRS5a6DulYfTjTrHmJKUf/view?usp=sharing

In case you don't want to download the file to see, below I pasted the bif file.

I was running this function gum.getPosterior(bn, evs= evidence, target=self.target), where
evidence = {'X1': 0}, self.target = X6

The graph looks like this

network "unnamedBN" {
// written by aGrUM 1.7.1.9
}

variable X5 {
   type discrete[2] {0, 1};
}

variable X3 {
   type discrete[2] {0, 1};
}

variable X6 {
   type discrete[2] {0, 1};
}

variable X4 {
   type discrete[2] {0, 1};
}

variable X1 {
   type discrete[2] {0, 1};
}

variable X7 {
   type discrete[2] {0, 1};
}

variable X0 {
   type discrete[2] {0, 1};
}

variable X2 {
   type discrete[2] {0, 1};
}

probability (X5 | X4, X0) {
   (0, 0) 0.147423 0.852577;
   (1, 0) 0.94209 0.0579096;
   (0, 1) 0.350282 0.649718;
   (1, 1) 0.0414797 0.95852;
}
probability (X3 | X2) {
   (0) 0.602883 0.397117;
   (1) 0.907101 0.0928988;
}
probability (X6 | X5, X2, X3) {
   (0, 0, 0) 0.402587 0.597413;
   (1, 0, 0) 0.707663 0.292337;
   (0, 1, 0) 0.777193 0.222807;
   (1, 1, 0) 0.522467 0.477533;
   (0, 0, 1) 0.990767 0.00923324;
   (1, 0, 1) 0.494996 0.505004;
   (0, 1, 1) 0.385113 0.614887;
   (1, 1, 1) 0.352897 0.647103;
}
probability (X4 | X2) {
   (0) 0.366486 0.633514;
   (1) 0.629139 0.370861;
}
probability (X1) {
   default 0.949951 0.050049;
}
probability (X7 | X5, X1, X3, X4) {
   (0, 0, 0, 0) 0.74636 0.25364;
   (1, 0, 0, 0) 0.450192 0.549808;
   (0, 1, 0, 0) 0.597444 0.402556;
   (1, 1, 0, 0) 0.195991 0.804009;
   (0, 0, 1, 0) 0.377 0.623;
   (1, 0, 1, 0) 0.98204 0.0179605;
   (0, 1, 1, 0) 0.609848 0.390152;
   (1, 1, 1, 0) 0.675516 0.324484;
   (0, 0, 0, 1) 0.435571 0.564429;
   (1, 0, 0, 1) 0.857639 0.142361;
   (0, 1, 0, 1) 0.77812 0.22188;
   (1, 1, 0, 1) 0.757985 0.242015;
   (0, 0, 1, 1) 0.323529 0.676471;
   (1, 0, 1, 1) 0.731697 0.268303;
   (0, 1, 1, 1) 0.937075 0.0629252;
   (1, 1, 1, 1) 0.259804 0.740196;
}
probability (X0) {
   default 0.437021 0.562979;
}
probability (X2 | X1) {
   (0) 0.423314 0.576686;
   (1) 0.759888 0.240112;
}

[phwuil]

Hi @kenneth-lee-ch , every things look fine here :

[kenneth-lee-ch]

Would you mind explaining what the error message is trying to say?

[kenneth-lee-ch]

I constantly run into this error

Here is another one bn file i save with evs = {'X2': 0} and target = 'X6

network "unnamedBN" {
// written by aGrUM 1.7.1.9
}

variable X3 {
   type discrete[2] {0, 1};
}

variable X1 {
   type discrete[2] {0, 1};
}

variable X5 {
   type discrete[2] {0, 1};
}

variable X7 {
   type discrete[2] {0, 1};
}

variable X2 {
   type discrete[2] {0, 1};
}

variable X4 {
   type discrete[2] {0, 1};
}

variable X0 {
   type discrete[2] {0, 1};
}

variable X6 {
   type discrete[2] {0, 1};
}

probability (X3) {
   default 0.877178 0.122822;
}
probability (X1 | X0) {
   (0) 0.840298 0.159702;
   (1) 0.403776 0.596224;
}
probability (X5 | X2, X4) {
   (0, 0) 0.882822 0.117178;
   (1, 0) 0.313253 0.686747;
   (0, 1) 0.0695691 0.930431;
   (1, 1) 0.675277 0.324723;
}
probability (X7 | X1, X4, X6, X2) {
   (0, 0, 0, 0) 0.851628 0.148372;
   (1, 0, 0, 0) 0.314685 0.685315;
   (0, 1, 0, 0) 0.129524 0.870476;
   (1, 1, 0, 0) 0.697051 0.302949;
   (0, 0, 1, 0) 0.657407 0.342593;
   (1, 0, 1, 0) 0.455373 0.544627;
   (0, 1, 1, 0) 0.973881 0.0261194;
   (1, 1, 1, 0) 0.315789 0.684211;
   (0, 0, 0, 1) 0.933962 0.0660377;
   (1, 0, 0, 1) 0.616667 0.383333;
   (0, 1, 0, 1) 0.797619 0.202381;
   (1, 1, 0, 1) 0.125 0.875;
   (0, 0, 1, 1) 0.554054 0.445946;
   (1, 0, 1, 1) 0.5 0.5;
   (0, 1, 1, 1) 0.325 0.675;
   (1, 1, 1, 1) 0.276923 0.723077;
}
probability (X2) {
   default 0.890603 0.109397;
}
probability (X4 | X3, X0) {
   (0, 0) 0.861335 0.138665;
   (1, 0) 0.311811 0.688189;
   (0, 1) 0.17906 0.82094;
   (1, 1) 0.723684 0.276316;
}
probability (X0) {
   default 0.727792 0.272208;
}
probability (X6 | X1, X4, X0, X2) {
   (0, 0, 0, 0) 0.844319 0.155681;
   (1, 0, 0, 0) 0.325464 0.674536;
   (0, 1, 0, 0) 0.186503 0.813497;
   (1, 1, 0, 0) 0.673759 0.326241;
   (0, 0, 1, 0) 0.570552 0.429448;
   (1, 0, 1, 0) 0.384298 0.615702;
   (0, 1, 1, 0) 0.725681 0.274319;
   (1, 1, 1, 0) 0.363874 0.636126;
   (0, 0, 0, 1) 0.836066 0.163934;
   (1, 0, 0, 1) 0.535211 0.464789;
   (0, 1, 0, 1) 0.655556 0.344444;
   (1, 1, 0, 1) 0.375 0.625;
   (0, 0, 1, 1) 0.461538 0.538462;
   (1, 0, 1, 1) 0.536585 0.463415;
   (0, 1, 1, 1) 0.337838 0.662162;
   (1, 1, 1, 1) 0.43299 0.56701;
}

[phwuil]

This error means that pyAgrum thinks that the type of your arguments are not among the possible ones. But I do not know where is the problem since here every thing is fine :

[phwuil]

[phwuil]

are you certain of the types and values of evidence and target ?

[kenneth-lee-ch]

I have this function

      def setPredVal(X, conditioning_set, sample_tuple, bn):
            """
                local function to set the prediction value to the dataframe for each sample
            """
            evidence = dict(zip(conditioning_set, sample_tuple))

            try:
                p = gum.getPosterior(bn, evs= evidence, target=self.target)
            except:
                print(type(sample_tuple[0]))
                print(type(sample_tuple[1]))
                print(type(sample_tuple))
                print(sample_tuple)
                print(type(evidence))
                print(evidence)
                print(type(self.target))
                gum.saveBN(bn, "problem_bn.bif")

this is the printed messages

<class 'numpy.int64'>
<class 'numpy.int64'>
<class 'tuple'>
(0, 1)
<class 'dict'>
{'X1': 0}
<class 'str'>

[kenneth-lee-ch]

I found the issue. Thanks for pointing out the potential issues with data type!