<?php
/**
 * DOCX Document Extractor.
 *
 * @package AISupportAssistant\Documents
 */

namespace AISupportAssistant\Documents;

use AISupportAssistant\Helpers\TextHelper;
use AISupportAssistant\Helpers\Logger;
use DOMDocument;
use ZipArchive;

if ( ! defined( 'ABSPATH' ) ) {
	exit;
}

/**
 * Class DocxExtractor
 */
class DocxExtractor implements ExtractorInterface {

	/**
	 * Extract text from a DOCX document.
	 *
	 * @param string $file_path Absolute path to the file.
	 * @return string Cleaned text.
	 */
	public function extract( string $file_path ): string {
		if ( ! file_exists( $file_path ) || ! class_exists( 'ZipArchive' ) ) {
			Logger::error( 'DocxExtractor failed: file does not exist or ZipArchive is missing.', array( 'path' => $file_path ) );
			return '';
		}

		$zip = new ZipArchive();
		if ( true !== $zip->open( $file_path ) ) {
			Logger::error( 'DocxExtractor failed: unable to open zip archive.', array( 'path' => $file_path ) );
			return '';
		}

		$xml_content = $zip->getFromName( 'word/document.xml' );
		$zip->close();

		if ( ! $xml_content ) {
			Logger::error( 'DocxExtractor failed: word/document.xml not found in zip.', array( 'path' => $file_path ) );
			return '';
		}

		// Parse XML using DOMDocument.
		$dom = new DOMDocument();
		
		// Disable external entity loading for security.
		$old_entity_loader = null;
		if ( function_exists( 'libxml_disable_entity_loader' ) && PHP_VERSION_ID < 80000 ) {
			// phpcs:ignore Generic.PHP.DeprecatedFunctions.Deprecated
			$old_entity_loader = libxml_disable_entity_loader( true );
		}
		
		libxml_use_internal_errors( true );
		$dom->loadXML( $xml_content );
		libxml_clear_errors();

		if ( $old_entity_loader !== null && PHP_VERSION_ID < 80000 ) {
			// phpcs:ignore Generic.PHP.DeprecatedFunctions.Deprecated
			libxml_disable_entity_loader( $old_entity_loader );
		}

		$paragraphs = $dom->getElementsByTagName( 'p' );
		$text_lines = array();

		foreach ( $paragraphs as $paragraph ) {
			$p_text = '';
			$texts  = $paragraph->getElementsByTagName( 't' );
			foreach ( $texts as $t ) {
				$p_text .= $t->nodeValue;
			}
			if ( '' !== trim( $p_text ) ) {
				$text_lines[] = $p_text;
			}
		}

		return TextHelper::clean_text( implode( "\n\n", $text_lines ) );
	}
}