///A tutorial about HMMs with silent states
#include <iostream>
#include <fstream>
#include <seqan/graph_algorithms.h>
#include <seqan/basic/basic_math.h>

using namespace seqan;

int main() {
///HMM creation
	typedef LogProb<> TProbability;
	typedef Dna TAlphabet;
	typedef Size<TAlphabet>::Type TSize;
	typedef Graph<Hmm<TAlphabet, TProbability> > THmm;
	typedef VertexDescriptor<THmm>::Type TVertexDescriptor;
	typedef EdgeDescriptor<THmm>::Type TEdgeDescriptor;
	
	Dna dnaA = Dna('A');
	Dna dnaC = Dna('C');
	Dna dnaG = Dna('G');
	Dna dnaT = Dna('T');

	THmm hmm;

///Begin state
	TVertexDescriptor begState = addVertex(hmm);
	assignBeginState(hmm, begState);

///Emission states
	TVertexDescriptor emitState1 = addVertex(hmm);
	emissionProbability(hmm, emitState1, dnaA) = 0.0;
	emissionProbability(hmm, emitState1, dnaC) = 0.8;
	emissionProbability(hmm, emitState1, dnaG) = 0.1;
	emissionProbability(hmm, emitState1, dnaT) = 0.1;

	TVertexDescriptor emitState2 = addVertex(hmm);
	emissionProbability(hmm, emitState2, dnaA) = 0.0;
	emissionProbability(hmm, emitState2, dnaC) = 0.2;
	emissionProbability(hmm, emitState2, dnaG) = 0.2;
	emissionProbability(hmm, emitState2, dnaT) = 0.6;

	TVertexDescriptor emitState3 = addVertex(hmm);
	emissionProbability(hmm, emitState3, dnaA) = 0.7;
	emissionProbability(hmm, emitState3, dnaC) = 0.1;
	emissionProbability(hmm, emitState3, dnaG) = 0.1;
	emissionProbability(hmm, emitState3, dnaT) = 0.1;

	TVertexDescriptor emitState4 = addVertex(hmm);
	emissionProbability(hmm, emitState4, dnaA) = 0.25;
	emissionProbability(hmm, emitState4, dnaC) = 0.25;
	emissionProbability(hmm, emitState4, dnaG) = 0.25;
	emissionProbability(hmm, emitState4, dnaT) = 0.25;

///Silent states (deletion states)
	TVertexDescriptor delState1 = addVertex(hmm, true);
	TVertexDescriptor delState2 = addVertex(hmm, true);
	TVertexDescriptor delState3 = addVertex(hmm, true);
	TVertexDescriptor delState4 = addVertex(hmm, true);

///End state
	TVertexDescriptor eState = addVertex(hmm);
	assignEndState(hmm, eState);

///Transitions
	addEdge(hmm, begState, emitState1, 0.5);
	addEdge(hmm, begState, delState1, 0.5);
	addEdge(hmm, emitState1, emitState2, 0.5);
	addEdge(hmm, emitState1, delState2, 0.5);
	addEdge(hmm, delState1, emitState2, 0.5);
	addEdge(hmm, delState1, delState2, 0.5);
	addEdge(hmm, emitState2, emitState3, 0.5);
	addEdge(hmm, emitState2, delState3, 0.5);
	addEdge(hmm, delState2, emitState3, 0.5);
	addEdge(hmm, delState2, delState3, 0.5);
	addEdge(hmm, emitState3, emitState4, 0.5);
	addEdge(hmm, emitState3, delState4, 0.5);
	addEdge(hmm, delState3, emitState4, 0.5);
	addEdge(hmm, delState3, delState4, 0.5);
	addEdge(hmm, emitState4, eState, 1.0);
	addEdge(hmm, delState4, eState, 1.0);

///Print the whole model
	::std::cout << hmm << ::std::endl;

///Viterbi algorithm
	String<Dna> sequence = "CA";
	String<TVertexDescriptor> path;
	TProbability p = viterbiAlgorithm(hmm, sequence, path);
	::std::cout << "Viterbi algorithm" << ::std::endl;
	::std::cout << "Probability of the best path: " << p << ::std::endl;
	::std::cout << "Sequence: " << ::std::endl;
	for(TSize i = 0; i<length(sequence); ++i) ::std::cout << sequence[i] << ',';
	::std::cout << ::std::endl;
	::std::cout << "State path: " << ::std::endl;
	for(TSize i = 0; i<length(path); ++i) {
		::std::cout << path[i];
		if (isSilent(hmm, path[i])) ::std::cout << " (Silent)";
		if (i < length(path) - 1) ::std::cout << ',';
	}
	::std::cout << ::std::endl;

///Forward algorithm
	::std::cout << "Forward algorithm" << ::std::endl;
	p = forwardAlgorithm(hmm, sequence);
	::std::cout << "Probability that the HMM generated the sequence: " << p << ::std::endl;

///Backward algorithm
	::std::cout << "Backward algorithm" << ::std::endl;
	p = backwardAlgorithm(hmm, sequence);
	::std::cout << "Probability that the HMM generated the sequence: " << p << ::std::endl;

	return 0;
}