// CharacterCodes.hpp


#if !defined(CHARACTER_CODES_HPP)
#define CHARACTER_CODES_HPP


namespace Cipres {

namespace SequenceFormats {


enum CharacterCode {
	// either
	GAP = -2,
	MISSING,
	// nucleotides
	ADENINE,
	CYTOSINE,
	GUANINE,
	THYMINE,
	URACIL,
	AMBIGUOUS_A_C,
	AMBIGUOUS_A_G,
	AMBIGUOUS_A_T,
	AMBIGUOUS_C_G,
	AMBIGUOUS_C_T,
	AMBIGUOUS_G_T,
	AMBIGUOUS_A_C_G,
	AMBIGUOUS_A_C_T,
	AMBIGUOUS_A_G_T,
	AMBIGUOUS_C_G_T,
	AMBIGUOUS_ANY,
	// amino acids
	ALANINE,
	CYSTEINE,
	ASPARTIC_ACID,
	GLUTAMIC_ACID,
	PHENYLALANINE,
	GLYCINE,
	HISTIDINE,
	ISOLEUCINE,
	LYSINE,
	LEUCINE,
	METHIONINE,
	ASPARAGINE,
	PROLINE,
	GLUTAMINE,
	ARGININE,
	SERINE,
	THREONINE,
	VALINE,
	TRYPTOPHAN,
	TYROSINE
};


inline int NucleotideSymbolToCode(char symbol) throw()
{
	static const char nucleotide_codes[256] = {
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, ADENINE, AMBIGUOUS_C_G_T, CYTOSINE, AMBIGUOUS_A_G_T, -1, -1, GUANINE, AMBIGUOUS_A_C_T, -1, -1, AMBIGUOUS_G_T, -1, AMBIGUOUS_A_C, AMBIGUOUS_ANY, -1,
		-1, -1, AMBIGUOUS_A_G, AMBIGUOUS_C_G, THYMINE, URACIL, AMBIGUOUS_A_C_G, AMBIGUOUS_A_T, AMBIGUOUS_ANY, AMBIGUOUS_C_T, -1, -1, -1, -1, -1, -1,
		-1, ADENINE, AMBIGUOUS_C_G_T, CYTOSINE, AMBIGUOUS_A_G_T, -1, -1, GUANINE, AMBIGUOUS_A_C_T, -1, -1, AMBIGUOUS_G_T, -1, AMBIGUOUS_A_C, AMBIGUOUS_ANY, -1,
		-1, -1, AMBIGUOUS_A_G, AMBIGUOUS_C_G, THYMINE, URACIL, AMBIGUOUS_A_C_G, AMBIGUOUS_A_T, AMBIGUOUS_ANY, AMBIGUOUS_C_T, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1
	};

	return nucleotide_codes[static_cast<unsigned int>(symbol)];
}

inline int AminoAcidSymbolToCode(char symbol) throw()
{
	static const char amino_acid_codes[256] = {
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, ALANINE, -1, CYSTEINE, ASPARTIC_ACID, GLUTAMIC_ACID, PHENYLALANINE, GLYCINE, HISTIDINE, ISOLEUCINE, -1, LYSINE, LEUCINE, METHIONINE, ASPARAGINE, -1,
		PROLINE, GLUTAMINE, ARGININE, SERINE, THREONINE, -1, VALINE, TRYPTOPHAN, -1, TYROSINE, -1, -1, -1, -1, -1, -1,
		-1, ALANINE, -1, CYSTEINE, ASPARTIC_ACID, GLUTAMIC_ACID, PHENYLALANINE, GLYCINE, HISTIDINE, ISOLEUCINE, -1, LYSINE, LEUCINE, METHIONINE, ASPARAGINE, -1,
		PROLINE, GLUTAMINE, ARGININE, SERINE, THREONINE, -1, VALINE, TRYPTOPHAN, -1, TYROSINE, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
		-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1
	};

	return amino_acid_codes[static_cast<unsigned int>(symbol)];
}


} // namespace SequenceFormats

} // namespace Cipres


#endif // CHARACTER_CODES_HPP