linguist-js/dist/src/program/parsing/parseGitattributes.js

Analyse the programming languages used in a folder or from raw content, using the same rules that GitHub Linguist does.

import { normPath } from '../fs/normalisedPath.js';
/**
 * Parses a gitattributes file.
 */
export default function parseGitattributes(content, folderRoot = '.') {
    const output = [];
    for (const rawLine of content.split('\n')) {
        const line = rawLine.replace(/#.*/, '').trim();
        if (!line)
            continue;
        const parts = line.split(/\s+/g);
        const fileGlob = parts[0];
        const relFileGlob = normPath(folderRoot, fileGlob);
        const attrParts = parts.slice(1);
        const isTrue = (str) => !str.startsWith('-') && !str.endsWith('=false');
        const isFalse = (str) => str.startsWith('-') || str.endsWith('=false');
        const trueParts = (str) => attrParts.filter(part => part.includes(str) && isTrue(part));
        const falseParts = (str) => attrParts.filter(part => part.includes(str) && isFalse(part));
        const hasTrueParts = (str) => trueParts(str).length > 0;
        const hasFalseParts = (str) => falseParts(str).length > 0;
        const boolOrNullVal = (str) => hasTrueParts(str) ? true : hasFalseParts(str) ? false : null;
        const attrs = {
            'generated': boolOrNullVal('linguist-generated'),
            'vendored': boolOrNullVal('linguist-vendored'),
            'documentation': boolOrNullVal('linguist-documentation'),
            'detectable': boolOrNullVal('linguist-detectable'),
            'binary': hasTrueParts('binary') || hasFalseParts('text') ? true : hasFalseParts('binary') || hasTrueParts('text') ? false : null,
            'language': trueParts('linguist-language').at(-1)?.split('=')[1] ?? null,
        };
        output.push({ glob: relFileGlob, attrs });
    }
    return output;
}