The function parseMolecule counts the number of atoms of each element contained in the molecule, in a given chemical formula represented by a string, and returns an object.
Sample output for formula H2O: {H: 2, O: 1}
A script by V.
Last active
April 30, 2017 21:10
-
-
Save rfprod/34f17716ebdde547d2486fc2dc468b84 to your computer and use it in GitHub Desktop.
Parse Molecule
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| parseMolecule = (formula) -> | |
| # console.log 'formula ' + formula | |
| elements = {} | |
| countElements = (input) -> | |
| # console.log 'count elements ' + input | |
| elts = input.match /[A-Z]([a-z])?(\d+)?/g | |
| for i in [0..elts.length - 1] by 1 | |
| elt = (elts[i].match /[A-Z]([a-z])?(?=\d+)?/)[0] | |
| count = elts[i].match /\d+/ | |
| count = if not count then 1 else (count.map (item) -> parseInt(item))[0] | |
| elements[elt] = if elements[elt] >= 0 then elements[elt] + count else count | |
| openBrackets = (input, closingBracket) -> | |
| # console.log 'open brackets, input ' + input | |
| output = '' | |
| multiplier = if (input.split closingBracket)[1] is '' then 1 else parseInt((input.split closingBracket)[1]) | |
| # console.log multiplier | |
| elts = input.match /[A-Z]([a-z])?(\d+)?/g | |
| for i in [0..elts.length - 1] by 1 | |
| elt = elts[i] | |
| eltName = (elt.match /^[A-Z]([a-z])?/)[0] | |
| atoms = if elt.match /\d+$/ then parseInt((elt.match /\d+$/)[0]) else 1 | |
| atoms *= multiplier | |
| output += eltName + atoms | |
| # console.log 'output ' + output | |
| formula = formula.replace input, output | |
| return output | |
| if /(\[|\]|\{|\}|\(|\))/.test formula | |
| # console.log 'extract inner round brackets first' | |
| formulaDecomp = formula.match /\([^\(\)]+\)\d*/g | |
| if formulaDecomp | |
| # console.log 'extract round, formulaDecomp ' + formulaDecomp | |
| for i in [0..formulaDecomp.length - 1] by 1 | |
| item = formulaDecomp[i] | |
| openBrackets item, ')' | |
| # console.log 'extract middle square brackets' | |
| formulaDecomp = formula.match /\[[^\[\]]+\]\d*/g | |
| if formulaDecomp | |
| # console.log 'extract square, formulaDecomp ' + formulaDecomp | |
| for i in [0..formulaDecomp.length - 1] by 1 | |
| item = formulaDecomp[i] | |
| openBrackets item, ']' | |
| # console.log 'extract outer curly brackets' | |
| formulaDecomp = formula.match /\{[^\[\]]+\}\d*/g | |
| if formulaDecomp | |
| # console.log 'extract curly, formulaDecomp ' + formulaDecomp | |
| for i in [0..formulaDecomp.length - 1] by 1 | |
| item = formulaDecomp[i] | |
| openBrackets item, '}' | |
| if not /(\[|\]|\{|\}|\(|\))/.test formula | |
| # console.log 'no brackets' | |
| countElements formula | |
| # console.log 'elements ' + elements | |
| return elements | |
| # TEST | |
| parseMolecule 'H2O' # {H: 2, O: 1}) | |
| parseMolecule 'Mg(OH)2' # {Mg: 1, O: 2, H: 2}) | |
| parseMolecule 'K4[ON(SO3)2]2' # {K: 4, O: 14, N: 2, S: 4}) | |
| parseMolecule 'As2{Be4C5[BCo3(CO2)3]2}4Cu5' # {As: 2, B: 8, Be: 16, C: 44, Cu: 5, Co: 24, O: 48} |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| function parseMolecule(formula) { | |
| // console.log('formula:', formula); | |
| let elements = {}; | |
| function countElements(input) { | |
| // console.log('count elements, input:', input); | |
| const elts = input.match(/[A-Z]([a-z])?(\d+)?/g); | |
| for (let i in elts) { | |
| const elt = elts[i].match(/[A-Z]([a-z])?(?=\d+)?/)[0]; | |
| let count = elts[i].match(/\d+/); | |
| count = (!count) ? 1 : count.map(Number)[0]; | |
| // console.log(elt, '|', count); | |
| if (elements.hasOwnProperty(elt)) { elements[elt] += count; } | |
| else { elements[elt] = count; } | |
| } | |
| // console.log('elements:', elements); | |
| } | |
| function openBrackets(input, closingBracket = ')') { | |
| let output = ''; | |
| const multiplier = parseInt(input.split(closingBracket)[1]) || 1; | |
| const elts = input.match(/[A-Z]([a-z])?(\d+)?/g); | |
| // console.log('elts: ',elts); | |
| for (let elt of elts) { | |
| const eltName = elt.match(/^[A-Z]([a-z])?/)[0]; | |
| // console.log('eltName:', eltName); | |
| let atoms = (elt.match(/\d+$/)) ? parseInt(elt.match(/\d+$/)[0]) : 1; | |
| // console.log('init atoms:', atoms); | |
| atoms *= multiplier; | |
| // console.log('resulting atoms:', atoms); | |
| output += eltName + atoms; | |
| } | |
| // console.log(output); | |
| formula = formula.replace(input, output); | |
| return output; | |
| }; | |
| if (/(\[|\]|\{|\}|\(|\))/.test(formula)) { | |
| // console.log('brackets detected, extract elements from brackets:', formula); | |
| // extract inner round brackets first | |
| let formulaDecomp = formula.match(new RegExp(/\([^\(\)]+\)\d*/, 'g')); | |
| // console.log('formulaDecomp, round brackets:', formulaDecomp); | |
| if (formulaDecomp) { | |
| for (let item of formulaDecomp) { | |
| //console.log('item:', item); | |
| openBrackets(item, ')'); | |
| // console.log(' >> formula:', formula); | |
| } | |
| } | |
| // extract middle square brackets | |
| formulaDecomp = formula.match(new RegExp(/\[[^\[\]]+\]\d*/, 'g')); | |
| // console.log('formulaDecomp, square brackets:', formulaDecomp); | |
| if (formulaDecomp) { | |
| for (let item of formulaDecomp) { | |
| openBrackets(item, ']'); | |
| // console.log(' >> formula:', formula); | |
| } | |
| } | |
| // extract outer curly brackets | |
| formulaDecomp = formula.match(new RegExp(/\{[^\[\]]+\}\d*/, 'g')); | |
| // console.log('formulaDecomp, curly brackets:', formulaDecomp); | |
| if (formulaDecomp) { | |
| for (let item of formulaDecomp) { | |
| openBrackets(item, '}'); | |
| // console.log(' >> formula:', formula); | |
| } | |
| } | |
| } | |
| if (!/(\[|\]|\{|\}|\(|\))/.test(formula)) { | |
| // console.log('no brackets, count elements:', formula); | |
| countElements(formula); | |
| } | |
| return elements; | |
| } | |
| // TEST | |
| parseMolecule('H2O'); // {H: 2, O: 1}) | |
| parseMolecule('Mg(OH)2'); // {Mg: 1, O: 2, H: 2}) | |
| parseMolecule('K4[ON(SO3)2]2'); // {K: 4, O: 14, N: 2, S: 4}) | |
| parseMolecule('As2{Be4C5[BCo3(CO2)3]2}4Cu5'); // {As: 2, B: 8, Be: 16, C: 44, Cu: 5, Co: 24, O: 48} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment