% Partial solution to Lab 9, the 'elegant' way (using regular expressions)
%
% Note : Extreme hackiness required because MATLAB is shithouse at string
% handling and requires strings to be stored in cell arrays.
%
% Note: Uses the containers.Map data structure, only available in MATLAB
% r2008b and better. (Why didn't MATLAB have an associative array data type
% in the first place?!)
%
% Li-aung Yip, liaung.yip@ieee.org

ElementSymbols = { 'H', 'He', 'Li', 'Be', 'O' };

% Mapping of element symbols to element weights.
ElementWeights = containers.Map({ 'H', 'He', 'Li', 'Be', 'O', 'Fe' },{ 1, 2, 3, 4, 16,26 });

Input = 'Fe2H2SO4Fe2O3';

% Step 1 : "Tokenise" the input string.
% That is, split the input into (element symbol)(number of atoms) pairs like so:
% Water:         'H20' -> 'H2', 'O'
% Sulphuric Acid 'H2SO4' -> 'H2', 'S', 'O4'
% Rust           'Fe2O3' -> 'Fe2', 'O3'
% Hexane         'CH3CH2CH2CH2CH2CH3' -> 'C', 'H3', 'C', 'H2', .... 'C','H3'.
%
% What's the pattern here? Each 'token' consists of an element symbol,
% optionally followed by a number. (An element symbol consists of one
% capital letter, optionally followed by some lowercase letters.)

TokenPattern = '[A-Z][a-z]*\d*';
Tokens = regexp(Input,TokenPattern,'match')

for Token = Tokens
    % Step 2 : Convert tokens like 'Fe2' to the symbol 'Fe' and the number 2.

    % Pull out the letters at the start
    ElementSymbol = regexp(Token,'[A-Z][a-z]*','match');
    ElementSymbol = ElementSymbol{1}{1}; % convert from cell array to text string.

    % Pull out the number at the end
    NumAtoms = regexp(Token,'\d*','match');
    if ( isempty(NumAtoms{1}) ) %If there were no numbers, assume one atom.
        NumAtoms = 1;
    else
        NumAtoms = str2num(NumAtoms{1}{1});
    end




    if ( ElementWeights.isKey(ElementSymbol) );
        ElementWeight = ElementWeights(ElementSymbol);
        TokenWeight = ElementWeight * NumAtoms;
            fprintf('Token %-6s | Symbol: %-3s | Number of: %-3i | Weight : %i * %i = %i\n',Token{1}, ElementSymbol,NumAtoms,NumAtoms,ElementWeight,NumAtoms*ElementWeight);
    else
        fprintf('I don''t know the weight of the element symbol %s\n',ElementSymbol);
    end

end









Sample Output:

Tokens =

    'Fe2'    'H2'    'S'    'O4'    'Fe2'    'O3'

Token Fe2    | Symbol: Fe  | Number of: 2   | Weight : 2 * 49 = 98
Token H2     | Symbol: H   | Number of: 2   | Weight : 2 * 1 = 2
I don't know the weight of the element symbol S
Token O4     | Symbol: O   | Number of: 4   | Weight : 4 * 16 = 64
Token Fe2    | Symbol: Fe  | Number of: 2   | Weight : 2 * 49 = 98
Token O3     | Symbol: O   | Number of: 3   | Weight : 3 * 16 = 48


Which is most of the way to a solution.