Most lexical analyzers are closely coupled with the language, the amount of code is relatively large. It's hard to pay attention to the essential principles of lexical analyzer.

In order to focus on the working principle of lexical analyzer , not to consider the small differences caused by different languages , an idea of making a lexer project that is completely decoupled from the language was born.

lexer realizes the decoupling of lexical analyzer and language through the following two JS files. lexer.js is the execution engine, {lang}-define.js is the language rule set, and their responsibilities are clearly separated:

• src/lexer.js is the core of the lexical analyzer, kept within 300 lines and split into ISR (Input Stream Reader) and DFA (Deterministic Finite Automaton) — extremely clear and easy to understand
• src/lang/{lang}-define.js is the language extension of the lexical analyzer, supporting different languages such as src/lang/c-define.js
  • ENUM_CONST — All enum values: token types, DFA state numbers, operator/symbol characters
  • CHARSET_CONST — Character set classifications: which characters are operators, symbols, keywords, etc.
  • DFA_STATE_CONST — DFA state constants (references values from ENUM_CONST)
  • tool — Utility functions: character type checks, token type inference, environment detection
  • flowModel — DFA state transition model: implements getNextState(ch, state) and records each step as a log entry

For the overall core execution flow, please refer to the Core Flowchart section

From inputting the character sequence to generating token after the analysis, lexer has complete steps for lexical analysis, and 11 token types for most language extensions

lexer supports different language extensions such as Python, Go, etc. How to make different language extensions, please check Contributions

• C ：A popular programming language，click here to see its lexical analysis
• SQL ：A popular database query language，click here to see its lexical analysis
• Goal ：A goal parser problem from leetCode ，click here to see its lexical analysis

The core mechanism of lexical analyzer is based on the state flow of DFA. For this reason, lexer records detailed state flow log to achieve the following requirements of you

• Debug mode
• Automatically generate DFA state flow diagram

After git clone command, no need for any dependencies, and no extra installation steps

Branch Note: The main branch is the primary branch, please use the main branch. Other branches are for feature development. All branches including main will have new feature iterations. Although the main branch has been tested multiple times, new features may still cause bugs. You can choose according to your needs, or download a more stable Release version

If you need use lexer in your project, such as code editor, etc.

npm install chain-lexer

var chainLexer = require('chain-lexer');
let lexer = chainLexer.cLexer;

let stream = "int a = 10;";
lexer.start(stream);
let parsedTokens = lexer.DFA.result.tokens;

lexer = chainLexer.sqlLexer;
stream = "select * from test where id >= 10;";
lexer.start(stream);
parsedTokens = lexer.DFA.result.tokens;

Import the package/{lang}-lexer.min.js file, then visit lexer variable to get the object of lexical analyzer，and visit lexer.DFA.result.tokens to get tokens

// 1. The code that needs lexical analysis
let stream = "int a = 10;";

// 2. Start lexical analysis
lexer.start(strem);

// 3. After the lexical analysis is done, get the generated tokens
let parsedTokens = lexer.DFA.result.tokens;

// 4. Do what you want to do
parsedTokens.forEach((token) => {
    // ... ...
});

The Provide state flow log part in features，visit flowModel.result.paths will get the detail logs of state flow inside lexer. The data format is as follows

[
    {
        state: 0, // now state
        ch: "a", // read char
        nextSstate: 2, // next state
        match: true, // is match
        end: false, // is last char
    },
    // ... ...
]

In order to preview the process of lexer in real time, to debug and test, there is a index.html file in the root directory of this project. Open it directly in your browser, and after entering the code will automatically output the Token generated after lexer analysis, as shown in the figure below

int a = 10;
int b =20;
int c = 20;

float f = 928.2332;
char b = 'b';

if(a == b){
    printf("Hello, World!");
}else if(b!=c){
    printf("Hello, World! Hello, World!");
}else{
    printf("Hello!");
}

or check the online website

• main: Primary branch
• develop: Development branch
• testing: Deprecated, no longer used
• v{x}: Version branch (e.g., v2 indicates version 2 branch), generally not used. Only during major version updates, a version branch is created. Code from develop branch is first merged into v{x} branch, and only after the version is fully tested and approved, it is merged into main branch

• test: Description related to testing
• perf: Description related to optimization
• feat: Description related to new features
• fix: Description related to bug fixes
• doc: Description related to documentation updates
• style: Description related to code formatting
• refactor: Description related to architectural refactoring

For example, if the lexer architecture is adjusted, the commit message should be refactor: refactor lexer

• Add more new features
• Add more extensions /src/lang/{lang}-define.js

The project is released with the version number of A-B-C，regarding release log, you can check the CHANGELOG or the release record

• A：Major upgrade
• B：Minor upgrade
• C：bug fix / features / ...

When a new version is ready, use the following commands to publish to npm:

git checkout main
git pull origin main
npm login
npm publish

Documents about source code development, project design, unit testing, automated testing, development specifications, and how to make extensions in different languages, please read source code explanation

If you have any problems or questions, please submit an issue