@arabold/docs-mcp-server 1.25.1 → 1.25.2

package/dist/index.js

@@ -709,7 +709,7 @@ function extractProtocol(urlOrPath) {
   }
 }
 const name = "@arabold/docs-mcp-server";
-const version = "1.25.0";
+const version = "1.25.1";
 const description = "MCP server for fetching and searching documentation";
 const type = "module";
 const bin = { "docs-mcp-server": "dist/index.js" };
@@ -4814,6 +4814,98 @@ class MarkdownMetadataExtractorMiddleware {
     await next();
   }
 }
+class HtmlNormalizationMiddleware {
+  async process(context, next) {
+    if (!context.dom) {
+      logger.debug(
+        `Skipping HTML normalization for ${context.source} - no DOM available`
+      );
+      await next();
+      return;
+    }
+    try {
+      logger.debug(`Normalizing HTML URLs and links for ${context.source}`);
+      const $ = context.dom;
+      const baseUrl = context.source;
+      this.normalizeImageUrls($, baseUrl);
+      this.normalizeLinks($, baseUrl);
+      logger.debug(`Successfully normalized HTML content for ${context.source}`);
+    } catch (error) {
+      logger.error(`❌ Failed to normalize HTML for ${context.source}: ${error}`);
+      context.errors.push(
+        error instanceof Error ? error : new Error(`HTML normalization failed: ${String(error)}`)
+      );
+    }
+    await next();
+  }
+  /**
+   * Normalizes image URLs by converting relative URLs to absolute URLs.
+   */
+  normalizeImageUrls($, baseUrl) {
+    $("img").each((_index, element) => {
+      const $img = $(element);
+      const src = $img.attr("src");
+      if (!src) return;
+      try {
+        new URL(src);
+      } catch {
+        try {
+          const absoluteUrl = new URL(src, baseUrl).href;
+          $img.attr("src", absoluteUrl);
+          logger.debug(`Converted relative image URL: ${src} → ${absoluteUrl}`);
+        } catch (error) {
+          logger.debug(`Failed to resolve relative image URL: ${src} - ${error}`);
+        }
+      }
+    });
+  }
+  /**
+   * Normalizes links by:
+   * - Converting relative URLs to absolute URLs
+   * - Unwrapping anchor links (preserving text content)
+   * - Unwrapping non-HTTP links (preserving text content)
+   */
+  normalizeLinks($, baseUrl) {
+    $("a").each((_index, element) => {
+      const $link = $(element);
+      const href = $link.attr("href");
+      if (!href) {
+        this.unwrapElement($, $link);
+        return;
+      }
+      if (href.startsWith("#")) {
+        logger.debug(`Removing anchor link: ${href}`);
+        this.unwrapElement($, $link);
+        return;
+      }
+      try {
+        const url = new URL(href);
+        if (url.protocol !== "http:" && url.protocol !== "https:") {
+          logger.debug(`Removing non-HTTP link: ${href}`);
+          this.unwrapElement($, $link);
+          return;
+        }
+      } catch {
+        try {
+          const absoluteUrl = new URL(href, baseUrl).href;
+          $link.attr("href", absoluteUrl);
+          logger.debug(`Converted relative link URL: ${href} → ${absoluteUrl}`);
+        } catch (error) {
+          logger.debug(`Failed to resolve relative link URL: ${href} - ${error}`);
+          this.unwrapElement($, $link);
+        }
+      }
+    });
+  }
+  /**
+   * Unwraps an element by replacing it with its HTML content.
+   * This preserves the inner HTML (including nested elements) while removing the wrapping tag.
+   */
+  unwrapElement(_$, $element) {
+    const htmlContent = $element.html() || $element.text();
+    $element.replaceWith(htmlContent);
+  }
+}
 function detectCharsetFromHtml(htmlContent) {
   const charsetMatch = htmlContent.match(
     /<meta\s+charset\s*=\s*["']?([^"'>\s]+)["']?[^>]*>/i
@@ -4937,6 +5029,7 @@ class HtmlPipeline extends BasePipeline {
       new HtmlMetadataExtractorMiddleware(),
       new HtmlLinkExtractorMiddleware(),
       new HtmlSanitizerMiddleware(),
+      new HtmlNormalizationMiddleware(),
       new HtmlToMarkdownMiddleware()
     ];
     const semanticSplitter = new SemanticMarkdownSplitter(