csv-parser.js

/**
 * CSV Parser Module
 * Handles loading and parsing CSV files with support for quoted values,
 * commas inside quotes, blank fields, and UTF-8 encoding.
 */
const CSVParser = (() => {
    'use strict';

    function parse(csvText) {
        const rows = [];
        const len = csvText.length;
        let i = 0;

        while (i < len) {
            const { row, nextIndex } = parseRow(csvText, i, len);
            if (row !== null) {
                rows.push(row);
            }
            i = nextIndex;
        }

        if (rows.length === 0) return { headers: [], data: [] };

        const headers = rows[0];
        const data = [];

        for (let r = 1; r < rows.length; r++) {
            const obj = {};
            for (let c = 0; c < headers.length; c++) {
                obj[headers[c].trim()] = (rows[r][c] || '').trim();
            }
            data.push(obj);
        }

        return { headers: headers.map(h => h.trim()), data };
    }

    function parseRow(text, start, len) {
        const fields = [];
        let i = start;

        if (i >= len) return { row: null, nextIndex: len };

        while (i < len) {
            if (text[i] === '"') {
                const result = parseQuotedField(text, i, len);
                fields.push(result.value);
                i = result.nextIndex;
            } else {
                const result = parseUnquotedField(text, i, len);
                fields.push(result.value);
                i = result.nextIndex;
            }

            if (i < len && text[i] === ',') {
                i++;
                if (i >= len) {
                    fields.push('');
                }
            } else {
                if (i < len && text[i] === '\r') i++;
                if (i < len && text[i] === '\n') i++;
                break;
            }
        }

        if (fields.length === 1 && fields[0] === '' && start < len &&
            (text[start] === '\n' || text[start] === '\r')) {
            return { row: null, nextIndex: i };
        }

        return { row: fields, nextIndex: i };
    }

    function parseQuotedField(text, start, len) {
        let value = '';
        let i = start + 1;

        while (i < len) {
            if (text[i] === '"') {
                if (i + 1 < len && text[i + 1] === '"') {
                    value += '"';
                    i += 2;
                } else {
                    i++;
                    break;
                }
            } else {
                value += text[i];
                i++;
            }
        }

        return { value, nextIndex: i };
    }

    function parseUnquotedField(text, start, len) {
        let end = start;
        while (end < len && text[end] !== ',' && text[end] !== '\n' && text[end] !== '\r') {
            end++;
        }
        return { value: text.substring(start, end), nextIndex: end };
    }

    async function loadCSV(url) {
        const response = await fetch(url);
        if (!response.ok) {
            throw new Error(`Failed to load CSV: ${response.status} ${response.statusText}`);
        }
        const text = await response.text();
        return parse(text);
    }

    return { parse, loadCSV };
})();