1.0.3 • Published 3 months ago

@pjsr/mathjs-lang v1.0.3

Weekly downloads
-
License
MIT
Repository
github
Last release
3 months ago

@pjsr/mathjs-lang

Small programming language for the 'mathjs' library, using a javascript sub-language. This way it becomes easy to use one of the existing text editors and highlight the code in javascript mode; small errors will occur due to the use of operators that do not exist in javascript.

If you like this library, if it helps you solve a problem, you have possibility, please offer a coffee.

Last changes!

 - updated 'mathjs' version to '^12.3.2'

 - removed 'to' and 'in' keywords in the parser, because cause conflit
   => use the 'unit' function instead, to work with units
 
 - removed ' transpose operator - use transpose function instead

How to use this library?

Create a node project, then install the library by running:

npm i @pjsr/mathjs-lang

In a code file, type the following:

import { Factory } from "@pjsr/mathjs-lang";
const factory = new Factory();
factory.math.import({
    logs: function() { 
        console.log(...arguments); 
    }
})
let code = 'logs("Hello world!")';
try {
    let parsed = factory.parse(code);
    console.log("\nPARSE result:");
    console.log( JSON.stringify(parsed) );
    console.log("\nRUN result:")
    factory.compile(parsed)()
}
catch(error) {
    console.log(error.message);
}

And the output will be:

PARSE result:
[[3,[5,"logs(\"Hello world!\")"],0,0],[0,0,0,0]]

RUN result:
Hello world!

The Factory.parse function generates an Array with a kind of intermediate code, which can be used with JSON for use in client/server mode.

Language statments

// variable declaration
let mydata = {
    name: "Pedro",
    age: 42
}

mydata.cold = true

// if statment
if(mydata.age < 40) {
    let msg = `${ mydata.name } is a new person`;
    logs( msg );
}
else {
    let msg = `${ mydata.name } is oldster`;
    logs( msg );
}

// for statment
for(let jj=40; jj < mydata.age; jj = jj + 1) {
    logs("value:", jj)
}

// switch statment
switch( mydata.age ) {
    case 40: logs("Age:", 40) break;
    case 41: logs("Age:", 41); break;
    case 42: logs("Age:", 42) break;
    default: logs("Unknow age")
}

// while statment
let tmp = 38
while(tmp < mydata.age) {
    logs("while tmp=", tmp)
    tmp = tmp + 2
}

// do ... while statment
do {
    logs("do while: tmp = ", tmp);
    tmp = tmp - 1;
} while(tmp >= mydata.age)

The result of the parse function is:

[[4,[3,"mydata"],[5,"{name: \"Pedro\", age: 42}"],0],[3,[5,"mydata.cold = true"],0,0],[2,[5,"mydata.age < 40"],0,[8,8]],[7,0,0,0],[4,[3,"msg"],[4,[[5,"mydata.name"],[3," is a new person"]]],0],[3,[5,"logs(msg)"],0,0],[8,0,0,0],[5,0,0,[8,12]],[7,0,0,0],[4,[3,"msg"],[4,[[5,"mydata.name"],[3," is oldster"]]],0],[3,[5,"logs(msg)"],0,0],[8,0,0,0],[7,0,0,[6,1]],[4,[3,"jj"],[5,"40"],0],[2,[5,"jj < mydata.age"],0,[8,18]],[3,[5,"logs(\"value:\", jj)"],0,0],[3,[5,"jj = jj + 1"],0,0],[5,0,0,[8,14]],[8,[6,1],0,0],[7,0,0,[6,2]],[3,[5,"mydata.age"],0,[6,3]],[2,[5,"40 != $3"],0,[8,26]],[2,[5,"41 != $3"],0,[8,28]],[2,[5,"42 != $3"],0,[8,30]],[5,0,0,[8,32]],[5,0,0,[8,33]],[3,[5,"logs(\"Age:\", 40)"],0,0],[5,0,0,[8,33]],[3,[5,"logs(\"Age:\", 41)"],0,0],[5,0,0,[8,33]],[3,[5,"logs(\"Age:\", 42)"],0,0],[5,0,0,[8,33]],[3,[5,"logs(\"Unknow age\")"],0,0],[8,[6,2],0,0],[4,[3,"tmp"],[5,"38"],0],[7,0,0,[6,4]],[2,[5,"tmp < mydata.age"],0,[8,40]],[3,[5,"logs(\"while tmp=\", tmp)"],0,0],[3,[5,"tmp = tmp + 2"],0,0],[5,0,0,[8,35]],[8,[6,4],0,0],[7,0,0,[6,5]],[3,[5,"logs(\"do while: tmp = \", tmp)"],0,0],[3,[5,"tmp = tmp - 1"],0,0],[2,[5,"not (tmp >= mydata.age)"],0,[8,41]],[8,[6,5],0,0],[0,0,0,0]]

The execution result is:

Pedro is oldster
value: 40
value: 41
Age: 42
while tmp= 38
while tmp= 40
do while: tmp =  42

The library makes use of the custom Scope with hierarchy that the mathjs library presents, so that whenever one of the previous code blocks is used, a new scope is created, which is destroyed at the end of the block execution.

This way, to create persistent variables, these can only be created in the root Scope, as is the example of 'mydata'.

The use of semicolons is almost always optional, except in 2 cases where it is mandatory:

  • break ;
  • return ;

The 'continue' statement is yet to be implemented;

The pattern string as a language object, not supported by 'mathjs', can only be used in 3 situations:

  • let v1 = ` ... ` (new variable declaration)
  • v2.name = ` ... ` (assignment of any)
  • return ` ... ` (return)

Language function

let hello = function(name) {
    let msg = concat("Hello ", name, "!");
    logs(msg)
    return [ log(42), msg];
}
logs(hello("Mike"))

Parse result:

[[4,[3,"hello"],[7,[["name"],[[4,[3,"msg"],[5,"concat(\"Hello \", name, \"!\")"],0],[3,[5,"logs(msg)"],0,0],[3,[5,"[log(42), msg]"],0,[0,null]],[5,0,0,[8,4]],[0,0,0,0]]]],0],[3,[5,"logs(hello(\"Mike\"))"],0,0],[0,0,0,0]]

Execution resukt:

Hello Mike!
[ 3.7376696182833684, 'Hello Mike!' ]

Another example:

let nd = {id: 1}
nd.make = function(id, name) { 
    return {id: id, name: name};
}
let nd1 = nd.make(log(10), "Cheese")
logs(nd, nd1)

Execution result:

{ id: 1, make: [Function: func] } { id: 2.302585092994046, name: 'Cheese' }

Creating a function can be in the following situations:

  • let v1 = function(...) (new variable declaration)
  • v2.name = function(...) (assignment)
  • return function(...); (return)

Functions, in addition to having access to their arguments, have access to all existing variables in the 'Scope' in which they are being executed, and only during their execution.

Functions that can be used

All functions present in the 'math' object of Factory are available for use. As a reference see mathjs help.

Operators

e
    : e OR e                {$$ = $1 + ' or ' + $3}
    | e XOR e               {$$ = $1 + ' xor ' + $3}
    | e AND e               {$$ = $1 + ' and ' + $3}
    | e '|' e               {$$ = $1 + ' | ' + $3}
    | e '^|' e              {$$ = $1 + ' ^| ' + $3}
    | e '&' e               {$$ = $1 + ' & ' + $3}
    | e '==' e              {$$ = $1 + ' == ' + $3}
    | e '!=' e              {$$ = $1 + ' != ' + $3}
    | e '>' e               {$$ = $1 + ' > ' + $3}
    | e '<' e               {$$ = $1 + ' < ' + $3}
    | e '>=' e              {$$ = $1 + ' >= ' + $3}
    | e '<=' e              {$$ = $1 + ' <= ' + $3}
    | e '>>>' e             {$$ = $1 + ' >>> ' + $3}
    | e '>>' e              {$$ = $1 + ' >> ' + $3}
    | e '<<' e              {$$ = $1 + ' << ' + $3}
    | e TO string           {$$ = $1 + ' to ' + $3}
    | e IN string           {$$ = "unit(" + $1 + ',' + $3 + ")";}
    | e ':' e %prec RANGE   {$$ = $1 + ' : ' + $3}
    | e '+' e               {$$ = $1 + ' + ' + $3;}
    | e '-' e               {$$ = $1 + ' - ' + $3;}
    | e '*' e               {$$ = $1 + ' * ' + $3;}
    | e '/' e               {$$ = $1 + ' / ' + $3;}
    | e '.*' e              {$$ = $1 + ' .* ' + $3;}
    | e './' e              {$$ = $1 + ' ./ ' + $3;}
    | e '%' e               {$$ = $1 + ' % ' + $3;}
    | e MOD e               {$$ = $1 + ' mod ' + $3;}
    | '-' e %prec UMINUS    {$$ = '-' + $2;}
    | '+' e %prec UPLUS     {$$ = '+' + $2;}
    | '~' e                 {$$ = '~ ' + $2;}
    | NOT e                 {$$ = 'not ' + $2;}
    | e '^' e               {$$ = $1 + ' ^ ' + $3;}
    | e '.^' e              {$$ = $1 + ' .^ ' + $3;}
    | e '!'                 {$$ = $1 + '!';}
    | e TRANSPOSE           {$$ = $1 + "'"; }
    | e '%'                 {$$ = $1 + '%';}
    | e '.' e               {$$ = $1 + '.' + $3;}
    | e '[' exprList ']'    {$$ = $1 + '[' + $3 + ']';}
    | e funcCall            {$$ = $1 + $2;}
    | object                {$$ = $1;}
    | matrix                {$$ = $1;}
    | '(' e ')'             {$$ = '(' + $2 + ')';}

    | HEX                   {$$ = yytext;}
    | OCT                   {$$ = yytext;}
    | BIN                   {$$ = yytext;}
    | NUMBER                {$$ = yytext;}
    | TRUE                  {$$ = 'true';}
    | FALSE                 {$$ = 'false';}
    | string                {$$ = $1;}
    | name                  {$$ = $1;}
    ;