Updated scanner logic to handle thumb VueJS components. Removed uneeded DLSiteWork fields.
All checks were successful
ci / build-test (push) Successful in 2m21s
ci / publish-image (push) Has been skipped

This commit is contained in:
2026-02-28 22:20:24 -05:00
parent ca7ffa1730
commit 704a6fc433
9 changed files with 316 additions and 40 deletions

View File

@@ -2,37 +2,26 @@
namespace JSMR.Infrastructure.Scanning.Models;
public class DLSiteHtmlDocument
public class DLSiteHtmlDocument(HtmlDocument document)
{
private readonly HtmlNodeCollection _workColumns;
private readonly HtmlNodeCollection _workColumnRights;
private readonly HtmlNodeCollection _workThumbs;
private readonly HtmlNodeCollection _workColumns = document.DocumentNode.SelectNodes("//dl[@class='work_1col']");
private readonly HtmlNodeCollection _workColumnRights = document.DocumentNode.SelectNodes("//td[contains(@class, 'work_1col_right')]");
private readonly HtmlNodeCollection _workThumbs = document.DocumentNode.SelectNodes("//div[@class='work_thumb']");
public HtmlNode PageTotalNode { get; } = document.DocumentNode.SelectNodes("//div[@class='page_total']/strong")[0];
public HtmlNode PageTotalNode { get; }
public DLSiteHtmlDocument(HtmlDocument document)
public DLSiteHtmlNode[] GetDLSiteNodes()
{
_workColumns = document.DocumentNode.SelectNodes("//dl[@class='work_1col']");
//_workColumnRights = document.DocumentNode.SelectNodes("//td[@class='work_1col_right']");
_workColumnRights = document.DocumentNode.SelectNodes("//td[contains(@class, 'work_1col_right')]");
_workThumbs = document.DocumentNode.SelectNodes("//div[@class='work_thumb']");
PageTotalNode = document.DocumentNode.SelectNodes("//div[@class='page_total']/strong")[0];
}
public List<DLSiteHtmlNode> GetDLSiteNodes()
{
var nodes = new List<DLSiteHtmlNode>();
List<DLSiteHtmlNode> nodes = [];
if (_workColumns.Count != _workColumnRights.Count || _workColumns.Count != _workThumbs.Count)
throw new Exception("Work column node counts do not match!");
for (int i = 0; i < _workColumns.Count; i++)
{
var node = new DLSiteHtmlNode(_workColumns[i], _workColumnRights[i], _workThumbs[i]);
DLSiteHtmlNode node = new(_workColumns[i], _workColumnRights[i], _workThumbs[i]);
nodes.Add(node);
}
return nodes;
return [.. nodes];
}
}

View File

@@ -19,7 +19,8 @@ public class DLSiteHtmlNode
public HtmlNode? SalesDateNode { get; private set; }
public HtmlNode DownloadsNode { get; private set; }
public HtmlNode? StarRatingNode { get; private set; }
public HtmlNode ImageNode { get; private set; }
public HtmlNode? ImageNode { get; private set; }
public HtmlNode? ThumbWithNgFilterBlockNode { get; private set; }
public HtmlNode[] GenreNodes { get; private set; }
public HtmlNode[] SearchTagNodes { get; private set; }
public HtmlNode[] CreatorNodes { get; private set; }
@@ -55,7 +56,8 @@ public class DLSiteHtmlNode
//InitializeSalesAndDownloadsNodes();
StarRatingNode = GetStarRatingNode();
ImageNode = GetImageNode();
ImageNode = TryGetImageNode();
ThumbWithNgFilterBlockNode = ThumbNode.SelectSingleNode(".//thumb-with-ng-filter-block");
}
private HtmlNode[] GetGenreNodes()
@@ -165,10 +167,13 @@ public class DLSiteHtmlNode
// }
//}
private HtmlNode GetImageNode()
private HtmlNode? TryGetImageNode()
{
HtmlNode linkNode = ThumbNode.SelectNodes(".//a")[0];
HtmlNode? linkNode = ThumbNode.SelectSingleNode(".//a");
return linkNode.SelectNodes(".//img")[0];
if (linkNode is null)
return null;
return linkNode.SelectSingleNode(".//img");
}
}