Pennington.Feeds.SitemapService

namespace Pennington.Feeds;

/// Generates sitemap XML for the /sitemap.xml endpoint. Uses AsyncLazy for lazy, thread-safe computation. When managed by FileWatchDependencyFactory, the instance is recreated on file changes — trusts IContentService for fresh metadata. Enumerates every DiscoverAsync result. For markdown sources, the parser is invoked so Date metadata (lastmod) and IsDraft filtering apply. For programmatic / Razor page sources, the route is emitted with no extra metadata — those content types are rarely dateable and forcing every programmatic generator to run at sitemap-build time would be expensive. Sitemap, search index, and llms.txt each answer a different question, so they intentionally run different filtering paths:
(this service) — every canonical HTML URL a crawler should index. Sourced from DiscoverAsync because Razor / programmatic routes with no TOC entry still need to appear. Per-page search: / llms: opt-outs are  honored here: those are search UX preferences, not SEO directives.
— enumerates GetIndexableEntriesAsync, which carries ExcludeFromSearch / ExcludeFromLlms flags sourced from search: / llms: front-matter.
Keep these two paths distinct — collapsing them would either leak search opt-outs into the sitemap (bad for SEO) or force every Razor/programmatic source to grow a TOC entry purely so it shows up in the sitemap.
public class SitemapService
{
    /// Initializes the service and prepares lazy sitemap generation driven by the provided builder.
    
public SitemapService(
    IEnumerable<IContentService> contentServices,
    LocalizationOptions localization,
    IEnumerable<IContentParser> parsers,
    SitemapBuilder builder);

    /// Returns the serialized sitemap XML, generating it on first access and caching the result.
    
public Task<string> GetSitemapXmlAsync();
}