worldhopper/lib/services/opds_parser.dart

import 'package:xml/xml.dart';
import 'package:worldhopper/config/constants.dart';
import 'package:worldhopper/models/enhanced_metadata.dart';
import 'package:worldhopper/models/opds_feed.dart';
import 'package:worldhopper/models/opds_entry.dart';
import 'package:worldhopper/models/opds_link.dart';
import 'package:worldhopper/models/opds_stream_link.dart';
import 'package:worldhopper/models/opds_server.dart';

/// Container for parsed feed and associated metadata
class ParsedFeed {
  final OPDSFeed feed;
  final Map<String, EnhancedMetadata> entryMetadata;

  const ParsedFeed(this.feed, this.entryMetadata);
}

/// Parser for OPDS 1.2 feeds with PSE (Page Streaming Extension) support
class OPDSParser {
  /// Parse an OPDS feed from XML string
  ParsedFeed parseFeed(String xmlContent, {OPDSServer? server}) {
    final document = XmlDocument.parse(xmlContent);
    final feedElement = document.findElements('feed').first;

    // Parse feed metadata
    final id = _getElementText(feedElement, 'id') ?? '';
    final title = _getElementText(feedElement, 'title') ?? 'Untitled';
    final subtitle = _getElementText(feedElement, 'subtitle');
    final icon = _getElementText(feedElement, 'icon');
    final updated = _parseDate(_getElementText(feedElement, 'updated'));

    // Parse author
    final authorElement = feedElement.findElements('author').firstOrNull;
    final author = authorElement != null
        ? _getElementText(authorElement, 'name')
        : null;

    // Parse feed links
    final feedLinks = feedElement
        .findElements('link')
        .map((element) => _parseLink(element, server: server))
        .where((link) => link != null)
        .cast<OPDSLink>()
        .toList();

    // Parse entries with metadata
    final entriesWithMetadata = feedElement
        .findElements('entry')
        .map((element) => _parseEntry(element, server))
        .toList();

    final entries = entriesWithMetadata.map((tuple) => tuple.$1).toList();
    final entryMetadata = Map.fromEntries(
      entriesWithMetadata.map((tuple) => MapEntry(tuple.$1.id, tuple.$2)),
    );

    // Parse OpenSearch elements (for pagination)
    int? totalResults;
    int? itemsPerPage;
    int? startIndex;

    try {
      final totalResultsText = feedElement
          .findElements('totalResults')
          .firstOrNull
          ?.innerText;
      if (totalResultsText != null) {
        totalResults = int.tryParse(totalResultsText);
      }

      final itemsPerPageText = feedElement
          .findElements('itemsPerPage')
          .firstOrNull
          ?.innerText;
      if (itemsPerPageText != null) {
        itemsPerPage = int.tryParse(itemsPerPageText);
      }

      final startIndexText = feedElement
          .findElements('startIndex')
          .firstOrNull
          ?.innerText;
      if (startIndexText != null) {
        startIndex = int.tryParse(startIndexText);
      }
    } catch (e) {
      // Ignore pagination parsing errors
    }

    final feed = OPDSFeed(
      id: id,
      title: title,
      subtitle: subtitle,
      icon: icon,
      updated: updated,
      author: author,
      links: feedLinks,
      entries: entries,
      totalResults: totalResults,
      itemsPerPage: itemsPerPage,
      startIndex: startIndex,
    );

    return ParsedFeed(feed, entryMetadata);
  }

  /// Parse an entry element with enhanced metadata
  (OPDSEntry, EnhancedMetadata) _parseEntry(XmlElement element, OPDSServer? server) {
    final id = _getElementText(element, 'id') ?? '';
    final title = _getElementText(element, 'title') ?? 'Untitled';
    final summary = _getElementText(element, 'summary');
    final content = _getElementText(element, 'content');
    final published = _parseDate(_getElementText(element, 'published'));
    final updated = _parseDate(_getElementText(element, 'updated'));

    // Parse authors
    final authors = element
        .findElements('author')
        .map((author) => _getElementText(author, 'name'))
        .where((name) => name != null)
        .cast<String>()
        .toList();

    // Parse categories
    final categories = element
        .findElements('category')
        .map((category) => category.getAttribute('term'))
        .where((term) => term != null)
        .cast<String>()
        .toList();

    // Parse links
    final links = element
        .findElements('link')
        .map((linkElement) => _parseLink(linkElement, server: server))
        .where((link) => link != null)
        .cast<OPDSLink>()
        .toList();

    // Parse PSE stream link
    final streamLink = _parseStreamLink(element, server);

    // Extract cover and thumbnail URLs and resolve them
    String? coverUrl;
    String? thumbnailUrl;

    for (final link in links) {
      if (link.rel.contains('image') || link.rel.contains('cover')) {
        final resolvedHref = server != null
            ? server.resolveUrl(link.href)
            : link.href;

        if (link.rel.contains('thumbnail')) {
          thumbnailUrl = resolvedHref;
        } else {
          coverUrl = resolvedHref;
        }
      }
    }

    final entry = OPDSEntry(
      id: id,
      title: title,
      summary: summary,
      content: content,
      published: published,
      updated: updated,
      authors: authors,
      categories: categories,
      links: links,
      streamLink: streamLink,
      coverUrl: coverUrl,
      thumbnailUrl: thumbnailUrl,
    );

    final metadata = _parseDCTermsMetadata(element);

    return (entry, metadata);
  }

  /// Parse a link element
  OPDSLink? _parseLink(XmlElement element, {OPDSServer? server}) {
    final rel = element.getAttribute('rel');
    final href = element.getAttribute('href');

    if (rel == null || href == null) return null;

    // Resolve href against server URL if server is provided
    // This ensures all links (acquisition, navigation, etc.) are absolute URLs
    final resolvedHref = server != null
        ? server.resolveUrl(href)
        : href;

    return OPDSLink(
      rel: rel,
      href: resolvedHref,  // Now absolute URL
      type: element.getAttribute('type'),
      title: element.getAttribute('title'),
    );
  }

  /// Parse a PSE stream link from an entry
  OPDSStreamLink? _parseStreamLink(XmlElement entryElement, OPDSServer? server) {
    try {
      // Find link with rel="http://vaemendis.net/opds-pse/stream"
      final linkElements = entryElement.findElements('link');

      for (final linkElement in linkElements) {
        final rel = linkElement.getAttribute('rel');

        if (rel == AppConstants.opdsPseStreamRel) {
          final href = linkElement.getAttribute('href');
          final type = linkElement.getAttribute('type');

          if (href == null || type == null) continue;

          // Parse PSE namespace attributes
          // Try different namespace prefixes
          int? pageCount;
          int? lastRead;
          DateTime? lastReadDate;

          // Try to find pse:count attribute
          final countAttr = linkElement.getAttribute('count',
                  namespace: AppConstants.opdsPseNamespace) ??
              linkElement.getAttribute('pse:count');

          if (countAttr != null) {
            pageCount = int.tryParse(countAttr);
          }

          // Try to find pse:lastRead attribute
          final lastReadAttr = linkElement.getAttribute('lastRead',
                  namespace: AppConstants.opdsPseNamespace) ??
              linkElement.getAttribute('pse:lastRead');

          if (lastReadAttr != null) {
            lastRead = int.tryParse(lastReadAttr);
          }

          // Try to find pse:lastReadDate attribute
          final lastReadDateAttr = linkElement.getAttribute('lastReadDate',
                  namespace: AppConstants.opdsPseNamespace) ??
              linkElement.getAttribute('pse:lastReadDate');

          if (lastReadDateAttr != null) {
            lastReadDate = _parseDate(lastReadDateAttr);
          }

          if (pageCount != null) {
            // Resolve href against server URL if server is provided
            // This matches the pattern used for cover/thumbnail URLs
            final resolvedHref = server != null
                ? server.resolveUrl(href)
                : href;

            return OPDSStreamLink(
              href: resolvedHref,
              type: type,
              pageCount: pageCount,
              lastRead: lastRead,
              lastReadDate: lastReadDate,
            );
          }
        }
      }
    } catch (e) {
      // Ignore PSE parsing errors
    }

    return null;
  }

  /// Get text content of a child element
  String? _getElementText(XmlElement parent, String elementName) {
    try {
      final element = parent.findElements(elementName).firstOrNull;
      return element?.innerText.trim();
    } catch (e) {
      return null;
    }
  }

  /// Parse an Atom date string
  DateTime? _parseDate(String? dateString) {
    if (dateString == null || dateString.isEmpty) return null;

    try {
      return DateTime.parse(dateString);
    } catch (e) {
      return null;
    }
  }

  /// Parse Dublin Core Terms metadata from entry element
  EnhancedMetadata _parseDCTermsMetadata(XmlElement element) {
    String? series;
    double? seriesPosition;
    String? publisher;
    String? language;
    String? isbn;

    try {
      // Try to find dcterms:isPartOf (series)
      final isPartOfElement = element
          .findElements('isPartOf', namespace: AppConstants.dctermsNamespace)
          .firstOrNull;
      if (isPartOfElement != null) {
        series = isPartOfElement.innerText.trim();
      }

      // Try schema:position or dcterms:position for series position
      final positionElement = element.findElements('position').firstOrNull;
      if (positionElement != null) {
        seriesPosition = double.tryParse(positionElement.innerText.trim());
      }

      // dcterms:publisher
      final publisherElement = element
          .findElements('publisher', namespace: AppConstants.dctermsNamespace)
          .firstOrNull;
      if (publisherElement != null) {
        publisher = publisherElement.innerText.trim();
      }

      // dcterms:language
      final languageElement = element
          .findElements('language', namespace: AppConstants.dctermsNamespace)
          .firstOrNull;
      if (languageElement != null) {
        language = languageElement.innerText.trim();
      }

      // dcterms:identifier - look for ISBN
      final identifierElements = element.findElements('identifier',
          namespace: AppConstants.dctermsNamespace);
      for (final idElement in identifierElements) {
        final text = idElement.innerText.trim().toUpperCase();
        if (text.contains('ISBN')) {
          // Extract ISBN: remove "ISBN:", "ISBN-13:", etc.
          isbn = text.replaceAll(RegExp(r'ISBN[- ]?(10|13)?:?\s*'), '').trim();
          break;
        }
      }
    } catch (e) {
      // Ignore dcterms parsing errors
    }

    return EnhancedMetadata(
      series: series,
      seriesPosition: seriesPosition,
      publisher: publisher,
      language: language,
      isbn: isbn,
    );
  }
}