Pennington.Pipeline.RenderedPage

namespace Pennington.Pipeline;

/// One page in the projected corpus: route + TOC entry + origin metadata, plus the post-pipeline HTML and a parsed AngleSharp element to aggregate over. Produced by ISiteProjection once per route per file-watch generation and shared across every corpus aggregator (search, llms.txt, link audit). Consumers pattern-match on Origin to recover front-matter / derived metadata where it exists. A nullOrigin means a Razor / programmatic page with no parsed item — the HTML is still available, just without front-matter context.No parsed DOM is retained. Holding an AngleSharp IElement keeps its whole owning document alive, so a corpus-sized cache of them costs roughly ten times what Html costs — enough to exhaust memory on large sites. Anything derived from the DOM is either captured here as value data (Sections) or re-parsed from Html on demand by the consumer that needs it, which keeps the parse transient.
public record RenderedPage
{
    /// Whether the page produced a content element; false for endpoint pages.
    
public bool HasContent { get; set; }


    /// Post-pipeline HTML of the selector-matched content element; empty for endpoint pages.
    
public string Html { get; set; }


    /// Origin information (markdown / endpoint), or null for Razor / programmatic pages.
    
public PageOrigin? Origin { get; set; }


    /// One page in the projected corpus: route + TOC entry + origin metadata, plus the post-pipeline HTML and a parsed AngleSharp element to aggregate over. Produced by ISiteProjection once per route per file-watch generation and shared across every corpus aggregator (search, llms.txt, link audit). Consumers pattern-match on Origin to recover front-matter / derived metadata where it exists. A nullOrigin means a Razor / programmatic page with no parsed item — the HTML is still available, just without front-matter context.No parsed DOM is retained. Holding an AngleSharp IElement keeps its whole owning document alive, so a corpus-sized cache of them costs roughly ten times what Html costs — enough to exhaust memory on large sites. Anything derived from the DOM is either captured here as value data (Sections) or re-parsed from Html on demand by the consumer that needs it, which keeps the parse transient.
    
public RenderedPage(ContentRoute Route, ContentTocItem Toc, PageOrigin? Origin, string Html, bool HasContent, IReadOnlyList<HeadingSection> Sections)
;

    /// Canonical route for the page.
    
public ContentRoute Route { get; set; }


    /// Heading-section split of the page content; empty for endpoint pages.
    
public IReadOnlyList<HeadingSection> Sections { get; set; }


    /// TOC entry that drove the page's inclusion.
    
public ContentTocItem Toc { get; set; }

}