vda-to-edifact-converter/js/edifact-validator.js

/**
 * EDIFACT Validator & Alternative Parser (Main Process only)
 * Uses ts-edifact library as a parallel layer alongside the custom DelforParser.
 *
 * IMPORTANT: This file runs in the Node.js Main Process only.
 * Do NOT include this file in index.html or use window.EDIBridge here.
 *
 * Exposed via IPC (see main.js):
 *   'validate-edifact'  → validateEdifact(content)
 *   'parse-edifact-lib' → parseWithLib(content)
 *
 * ts-edifact API:
 *   const { Reader } = require('ts-edifact');
 *   const reader = new Reader();       // instanciate per call (stateful)
 *   const segments = reader.parse(str); // returns { name, elements: string[][] }[]
 */

'use strict';

/**
 * Parse EDIFACT content using ts-edifact Reader.
 * Returns a clean segment array for comparison with the custom DelforParser.
 *
 * Segment format: { name: 'UNH', elements: [['1'], ['DELFOR','D','04A','UN']] }
 *
 * @param {string} content - Raw EDIFACT content (string)
 * @returns {{ segments: Array, segmentCount: number, error?: string }}
 */
function parseWithLib(content) {
    try {
        const { Reader, NullValidator, Parser } = require('ts-edifact');

        // Normalize line endings
        const normalized = content.replace(/\r\n/g, '').replace(/\r/g, '').trim();

        // Each Reader instance is stateful — create a new one per call
        const reader = new Reader();

        // --- FIX: Disable strict segment validation & Enable UNOC ---
        const validator = new NullValidator();
        if (typeof validator.define !== 'function') validator.define = () => { };

        reader.validator = validator;
        reader.parser = new Parser(validator);
        reader.defined = true;
        reader.encoding('UNOC'); // Allow lowercase and special chars

        // Re-setup Reader callbacks for the new parser
        const result = reader.result;
        let elements = reader.elements;
        let components = reader.components;
        reader.parser.onOpenSegment = function (segment) {
            elements = [];
            result.push({ name: segment, elements: elements });
        };
        reader.parser.onElement = function () {
            components = [];
            elements.push(components);
        };
        reader.parser.onComponent = function (value) {
            components.push(value);
        };

        const segments = reader.parse(normalized);

        return {
            segments,
            segmentCount: segments.length
        };

    } catch (e) {
        return {
            segments: [],
            segmentCount: 0,
            error: e.message
        };
    }
}

/**
 * Validate EDIFACT content using ts-edifact library + structural checks.
 *
 * Performs:
 *   1. Syntax parsing via ts-edifact Reader (catches malformed EDI)
 *   2. Structural checks: UNB/UNZ envelope, UNH/UNT balance, BGM presence
 *   3. UNT segment count verification
 *   4. Message type & version detection
 *
 * @param {string} content - Raw EDIFACT content
 * @returns {{ valid: boolean, errors: string[], warnings: string[], stats: object }}
 */
function validateEdifact(content) {
    const errors = [];
    const warnings = [];
    const stats = {};

    try {
        const { Reader, NullValidator, Parser } = require('ts-edifact');

        // --- Step 1: Parse with ts-edifact ---
        let segments = null;
        try {
            const normalized = content.replace(/\r\n/g, '').replace(/\r/g, '').trim();
            const reader = new Reader();

            // --- FIX: Disable strict segment validation & Enable UNOC ---
            const validator = new NullValidator();
            if (typeof validator.define !== 'function') validator.define = () => { };

            reader.validator = validator;
            reader.parser = new Parser(validator);
            reader.defined = true;
            reader.encoding('UNOC');

            // Re-setup Reader callbacks
            const result = reader.result;
            let elements = reader.elements;
            let components = reader.components;
            reader.parser.onOpenSegment = function (segment) {
                elements = [];
                result.push({ name: segment, elements: elements });
            };
            reader.parser.onElement = function () {
                components = [];
                elements.push(components);
            };
            reader.parser.onComponent = function (value) {
                components.push(value);
            };

            segments = reader.parse(normalized);
        } catch (parseErr) {
            errors.push('Syntaxfehler: ' + parseErr.message);
        }

        if (!segments || !Array.isArray(segments)) {
            return { valid: errors.length === 0, errors, warnings, stats };
        }

        // Build a flat tag list for quick lookups
        const tags = segments.map(s => s.name);
        stats.segmentCount = segments.length;
        stats.allSegments = tags;

        // --- Step 2: Interchange envelope (UNB / UNZ) ---
        if (!tags.includes('UNB')) {
            warnings.push('Kein UNB-Segment (Interchange-Header) gefunden.');
        } else {
            const unb = segments.find(s => s.name === 'UNB');
            if (unb && unb.elements[1]) stats.sender = unb.elements[1][0] || '';
            if (unb && unb.elements[2]) stats.receiver = unb.elements[2][0] || '';
        }
        if (!tags.includes('UNZ')) warnings.push('Kein UNZ-Segment (Interchange-Trailer) gefunden.');

        // --- Step 3: Message envelope (UNH / UNT) ---
        const unhCount = tags.filter(t => t === 'UNH').length;
        const untCount = tags.filter(t => t === 'UNT').length;
        stats.messageCount = unhCount;

        if (unhCount === 0) {
            errors.push('Kein UNH-Segment (Message-Header) gefunden.');
        } else {
            // Detect message type from first UNH+<ref>+<type>:<ver>:<release>:<ctrl>
            const unh = segments.find(s => s.name === 'UNH');
            if (unh && unh.elements[1]) {
                stats.messageType = unh.elements[1][0] || '';
                stats.messageVersion = (unh.elements[1][1] || '') + (unh.elements[1][2] || '');
                stats.controlRef = unh.elements[0]?.[0] || '';
            }
        }

        if (untCount === 0) {
            errors.push('Kein UNT-Segment (Message-Trailer) gefunden.');
        }

        if (unhCount !== untCount) {
            errors.push(`UNH/UNT-Anzahl stimmt nicht überein: ${unhCount} × UNH vs ${untCount} × UNT.`);
        }

        // --- Step 4: UNT segment count check ---
        // UNT+<segCount>+<msgRef>' — segCount includes UNH and UNT itself
        const unt = segments.find(s => s.name === 'UNT');
        if (unt && unt.elements[0]) {
            const declared = parseInt(unt.elements[0][0]);
            // Find the slice from UNH to UNT (inclusive)
            const unhIdx = segments.findIndex(s => s.name === 'UNH');
            const untIdx = segments.findIndex(s => s.name === 'UNT');
            if (!isNaN(declared) && unhIdx >= 0 && untIdx >= 0) {
                const actual = untIdx - unhIdx + 1;
                if (declared !== actual) {
                    warnings.push(
                        `UNT deklariert ${declared} Segmente, gezählt (UNH→UNT): ${actual}.`
                    );
                }
            }
        }

        // --- Step 5: BGM check ---
        if (!tags.includes('BGM')) {
            warnings.push('Kein BGM-Segment (Nachrichtenbeginn) gefunden.');
        } else {
            const bgm = segments.find(s => s.name === 'BGM');
            if (bgm && bgm.elements[0]) stats.documentCode = bgm.elements[0][0] || '';
        }

    } catch (e) {
        errors.push('Unerwarteter Validierungsfehler: ' + e.message);
    }

    return {
        valid: errors.length === 0,
        errors,
        warnings,
        stats
    };
}

module.exports = { validateEdifact, parseWithLib };