einvoice/ts/formats/validation/schematron.validator.ts

import * as plugins from '../../plugins.js';
import * as SaxonJS from 'saxon-js';
import type { ValidationResult } from './validation.types.js';

/**
 * Schematron validation options
 */
export interface SchematronOptions {
  phase?: string; // Schematron phase to activate
  parameters?: Record<string, any>; // Parameters to pass to Schematron
  includeWarnings?: boolean; // Include warning-level messages
  maxErrors?: number; // Maximum errors before stopping
}

/**
 * Schematron validation engine using Saxon-JS
 * Provides official standards validation through Schematron rules
 */
export class SchematronValidator {
  private compiledStylesheet: any;
  private schematronRules: string;
  private isCompiled: boolean = false;
  
  constructor(schematronRules?: string) {
    this.schematronRules = schematronRules || '';
  }
  
  /**
   * Load Schematron rules from file or string
   */
  public async loadSchematron(source: string, isFilePath: boolean = true): Promise<void> {
    if (isFilePath) {
      // Load from file
      const smartfile = await import('@push.rocks/smartfile');
      this.schematronRules = await smartfile.SmartFile.fromFilePath(source).then(f => f.contentBuffer.toString());
    } else {
      // Use provided string
      this.schematronRules = source;
    }
    
    // Reset compilation state
    this.isCompiled = false;
  }
  
  /**
   * Compile Schematron to XSLT using ISO Schematron skeleton
   */
  private async compileSchematron(): Promise<void> {
    if (this.isCompiled) return;
    
    // The Schematron to XSLT transformation requires the ISO Schematron skeleton
    // For now, we'll use a simplified approach with direct XSLT generation
    // In production, we would use the official ISO Schematron skeleton XSLTs
    
    try {
      // Convert Schematron to XSLT
      // This is a simplified version - in production we'd use the full ISO skeleton
      const xslt = this.generateXSLTFromSchematron(this.schematronRules);
      
      // Compile the XSLT with Saxon-JS
      this.compiledStylesheet = await SaxonJS.compile({
        stylesheetText: xslt,
        warnings: 'silent'
      });
      
      this.isCompiled = true;
    } catch (error) {
      throw new Error(`Failed to compile Schematron: ${error.message}`);
    }
  }
  
  /**
   * Validate an XML document against loaded Schematron rules
   */
  public async validate(
    xmlContent: string,
    options: SchematronOptions = {}
  ): Promise<ValidationResult[]> {
    if (!this.schematronRules) {
      throw new Error('No Schematron rules loaded');
    }
    
    // Ensure Schematron is compiled
    await this.compileSchematron();
    
    const results: ValidationResult[] = [];
    
    try {
      // Transform the XML with the compiled Schematron XSLT
      const transformResult = await SaxonJS.transform({
        stylesheetInternal: this.compiledStylesheet,
        sourceText: xmlContent,
        destination: 'serialized',
        stylesheetParams: options.parameters || {}
      });
      
      // Parse the SVRL (Schematron Validation Report Language) output
      results.push(...this.parseSVRL(transformResult.principalResult));
      
      // Apply options filters
      if (!options.includeWarnings) {
        return results.filter(r => r.severity !== 'warning');
      }
      
      if (options.maxErrors && results.filter(r => r.severity === 'error').length > options.maxErrors) {
        return results.slice(0, options.maxErrors);
      }
      
      return results;
    } catch (error) {
      results.push({
        ruleId: 'SCHEMATRON-ERROR',
        source: 'SCHEMATRON',
        severity: 'error',
        message: `Schematron validation failed: ${error.message}`,
        btReference: undefined,
        bgReference: undefined
      });
      return results;
    }
  }
  
  /**
   * Parse SVRL output to ValidationResult array
   */
  private parseSVRL(svrlXml: string): ValidationResult[] {
    const results: ValidationResult[] = [];
    
    // Parse SVRL XML
    const parser = new plugins.xmldom.DOMParser();
    const doc = parser.parseFromString(svrlXml, 'text/xml');
    
    // Get all failed assertions and successful reports
    const failedAsserts = doc.getElementsByTagName('svrl:failed-assert');
    const successfulReports = doc.getElementsByTagName('svrl:successful-report');
    
    // Process failed assertions (these are errors)
    for (let i = 0; i < failedAsserts.length; i++) {
      const assert = failedAsserts[i];
      const result = this.extractValidationResult(assert, 'error');
      if (result) results.push(result);
    }
    
    // Process successful reports (these can be warnings or info)
    for (let i = 0; i < successfulReports.length; i++) {
      const report = successfulReports[i];
      const result = this.extractValidationResult(report, 'warning');
      if (result) results.push(result);
    }
    
    return results;
  }
  
  /**
   * Extract ValidationResult from SVRL element
   */
  private extractValidationResult(
    element: Element,
    defaultSeverity: 'error' | 'warning'
  ): ValidationResult | null {
    const text = element.getElementsByTagName('svrl:text')[0]?.textContent || '';
    const location = element.getAttribute('location') || undefined;
    const test = element.getAttribute('test') || '';
    const id = element.getAttribute('id') || element.getAttribute('role') || 'UNKNOWN';
    const flag = element.getAttribute('flag') || defaultSeverity;
    
    // Determine severity from flag attribute
    let severity: 'error' | 'warning' | 'info' = defaultSeverity;
    if (flag.toLowerCase().includes('fatal') || flag.toLowerCase().includes('error')) {
      severity = 'error';
    } else if (flag.toLowerCase().includes('warning')) {
      severity = 'warning';
    } else if (flag.toLowerCase().includes('info')) {
      severity = 'info';
    }
    
    // Extract BT/BG references if present
    const btMatch = text.match(/\[BT-(\d+)\]/);
    const bgMatch = text.match(/\[BG-(\d+)\]/);
    
    return {
      ruleId: id,
      source: 'EN16931',
      severity,
      message: text,
      syntaxPath: location,
      btReference: btMatch ? `BT-${btMatch[1]}` : undefined,
      bgReference: bgMatch ? `BG-${bgMatch[1]}` : undefined,
      profile: 'EN16931'
    };
  }
  
  /**
   * Generate simplified XSLT from Schematron
   * This is a placeholder - in production, use ISO Schematron skeleton
   */
  private generateXSLTFromSchematron(schematron: string): string {
    // This is a simplified transformation
    // In production, we would use the official ISO Schematron skeleton XSLTs
    // (iso_schematron_skeleton.xsl, iso_svrl_for_xslt2.xsl, etc.)
    
    // For now, return a basic XSLT that creates SVRL output
    return `<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet version="3.0" 
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
    xmlns:svrl="http://purl.oclc.org/dsdl/svrl">
    
  <xsl:output method="xml" indent="yes"/>
  
  <xsl:template match="/">
    <svrl:schematron-output>
      <!-- This is a placeholder transformation -->
      <!-- Real implementation would process Schematron patterns and rules -->
      <svrl:active-pattern>
        <xsl:attribute name="document">
          <xsl:value-of select="base-uri(/)"/>
        </xsl:attribute>
      </svrl:active-pattern>
    </svrl:schematron-output>
  </xsl:template>
</xsl:stylesheet>`;
  }
  
  /**
   * Check if validator has rules loaded
   */
  public hasRules(): boolean {
    return !!this.schematronRules;
  }
  
  /**
   * Get list of available phases from Schematron
   */
  public async getPhases(): Promise<string[]> {
    if (!this.schematronRules) return [];
    
    const parser = new plugins.xmldom.DOMParser();
    const doc = parser.parseFromString(this.schematronRules, 'text/xml');
    const phases = doc.getElementsByTagName('sch:phase');
    
    const phaseNames: string[] = [];
    for (let i = 0; i < phases.length; i++) {
      const id = phases[i].getAttribute('id');
      if (id) phaseNames.push(id);
    }
    
    return phaseNames;
  }
  
  /**
   * Validate with specific phase activated
   */
  public async validateWithPhase(
    xmlContent: string,
    phase: string,
    options: SchematronOptions = {}
  ): Promise<ValidationResult[]> {
    return this.validate(xmlContent, { ...options, phase });
  }
}

/**
 * Factory function to create validator with standard Schematron packs
 */
export async function createStandardValidator(
  standard: 'EN16931' | 'XRECHNUNG' | 'PEPPOL' | 'FACTURX'
): Promise<SchematronValidator> {
  const validator = new SchematronValidator();
  
  // Load appropriate Schematron based on standard
  // These paths would point to actual Schematron files in production
  switch (standard) {
    case 'EN16931':
      // Would load from ConnectingEurope/eInvoicing-EN16931
      await validator.loadSchematron('assets/schematron/en16931/EN16931-UBL-validation.sch');
      break;
    case 'XRECHNUNG':
      // Would load from itplr-kosit/xrechnung-schematron
      await validator.loadSchematron('assets/schematron/xrechnung/XRechnung-UBL-validation.sch');
      break;
    case 'PEPPOL':
      // Would load from OpenPEPPOL/peppol-bis-invoice-3
      await validator.loadSchematron('assets/schematron/peppol/PEPPOL-EN16931-UBL.sch');
      break;
    case 'FACTURX':
      // Would load from Factur-X specific Schematron
      await validator.loadSchematron('assets/schematron/facturx/Factur-X-EN16931-validation.sch');
      break;
  }
  
  return validator;
}

/**
 * Hybrid validator that combines TypeScript and Schematron validation
 */
export class HybridValidator {
  private schematronValidator: SchematronValidator;
  private tsValidators: Array<{ validate: (xml: string) => ValidationResult[] }> = [];
  
  constructor(schematronValidator?: SchematronValidator) {
    this.schematronValidator = schematronValidator || new SchematronValidator();
  }
  
  /**
   * Add a TypeScript validator to the pipeline
   */
  public addTSValidator(validator: { validate: (xml: string) => ValidationResult[] }): void {
    this.tsValidators.push(validator);
  }
  
  /**
   * Run all validators and merge results
   */
  public async validate(
    xmlContent: string,
    options: SchematronOptions = {}
  ): Promise<ValidationResult[]> {
    const results: ValidationResult[] = [];
    
    // Run TypeScript validators first (faster, better UX)
    for (const validator of this.tsValidators) {
      try {
        results.push(...validator.validate(xmlContent));
      } catch (error) {
        console.warn(`TS validator failed: ${error.message}`);
      }
    }
    
    // Run Schematron validation if available
    if (this.schematronValidator.hasRules()) {
      try {
        const schematronResults = await this.schematronValidator.validate(xmlContent, options);
        results.push(...schematronResults);
      } catch (error) {
        console.warn(`Schematron validation failed: ${error.message}`);
      }
    }
    
    // Deduplicate results by ruleId
    const seen = new Set<string>();
    return results.filter(r => {
      if (seen.has(r.ruleId)) return false;
      seen.add(r.ruleId);
      return true;
    });
  }
}