@arabold/docs-mcp-server 1.12.4 → 1.14.0

package/dist/{DocumentManagementService-BupnR1eC.js → DocumentManagementService-BZ_ZZgPI.js}

@@ -10,6 +10,7 @@ import { VirtualConsole, JSDOM } from "jsdom";
 import { chromium } from "playwright";
 import { gfm } from "@joplin/turndown-plugin-gfm";
 import TurndownService from "turndown";
+import { TextDecoder } from "node:util";
 import { URL as URL$1, fileURLToPath } from "node:url";
 import * as semver from "semver";
 import semver__default from "semver";
@@ -168,6 +169,49 @@ const FETCHER_BASE_DELAY = 1e3;
 const SPLITTER_MIN_CHUNK_SIZE = 500;
 const SPLITTER_PREFERRED_CHUNK_SIZE = 1500;
 const SPLITTER_MAX_CHUNK_SIZE = 5e3;
+const EMBEDDING_BATCH_SIZE = 100;
+class MimeTypeUtils {
+  /**
+   * Parses a Content-Type header string into its MIME type and charset.
+   * @param contentTypeHeader The Content-Type header string (e.g., "text/html; charset=utf-8").
+   * @returns A ParsedContentType object, or a default if parsing fails.
+   */
+  static parseContentType(contentTypeHeader) {
+    if (!contentTypeHeader) {
+      return { mimeType: "application/octet-stream" };
+    }
+    const parts = contentTypeHeader.split(";").map((part) => part.trim());
+    const mimeType = parts[0].toLowerCase();
+    let charset;
+    for (let i = 1; i < parts.length; i++) {
+      const param = parts[i];
+      if (param.toLowerCase().startsWith("charset=")) {
+        charset = param.substring("charset=".length).toLowerCase();
+        break;
+      }
+    }
+    return { mimeType, charset };
+  }
+  /**
+   * Checks if a MIME type represents HTML content.
+   */
+  static isHtml(mimeType) {
+    return mimeType === "text/html" || mimeType === "application/xhtml+xml";
+  }
+  /**
+   * Checks if a MIME type represents Markdown content.
+   */
+  static isMarkdown(mimeType) {
+    return mimeType === "text/markdown" || mimeType === "text/x-markdown";
+  }
+  /**
+   * Checks if a MIME type represents plain text content.
+   */
+  static isText(mimeType) {
+    return mimeType.startsWith("text/");
+  }
+  // Extend with more helpers as needed (isJson, isXml, isPdf, etc.)
+}
 class FingerprintGenerator {
   headerGenerator;
   /**
@@ -245,11 +289,15 @@ class HttpFetcher {
           maxRedirects: followRedirects ? 5 : 0
         };
         const response = await axios.get(source, config);
+        const contentTypeHeader = response.headers["content-type"];
+        const { mimeType, charset } = MimeTypeUtils.parseContentType(contentTypeHeader);
+        const contentEncoding = response.headers["content-encoding"];
         return {
           content: response.data,
-          mimeType: response.headers["content-type"] || "application/octet-stream",
-          source,
-          encoding: response.headers["content-encoding"]
+          mimeType,
+          charset,
+          encoding: contentEncoding,
+          source
         };
       } catch (error) {
         const axiosError = error;
@@ -322,55 +370,11 @@ class FileFetcher {
     }
   }
 }
-class ContentProcessingPipeline {
-  middleware;
-  /**
-   * Creates an instance of ContentProcessingPipeline.
-   * @param middleware An array of middleware instances to execute in order.
-   */
-  constructor(middleware) {
-    this.middleware = middleware;
-  }
-  /**
-   * Executes the middleware pipeline with the given initial context.
-   * @param initialContext The starting context for the pipeline.
-   * @returns A promise that resolves with the final context after all middleware have executed.
-   */
-  async run(initialContext) {
-    let index = -1;
-    const dispatch = async (i) => {
-      if (i <= index) {
-        throw new Error("next() called multiple times");
-      }
-      index = i;
-      const mw = this.middleware[i];
-      if (!mw) {
-        return;
-      }
-      const next = dispatch.bind(null, i + 1);
-      try {
-        await mw.process(initialContext, next);
-      } catch (error) {
-        initialContext.errors.push(
-          error instanceof Error ? error : new Error(String(error))
-        );
-        logger.warn(`Error in middleware pipeline: ${error}`);
-      }
-    };
-    await dispatch(0);
-    return initialContext;
-  }
-}
 class HtmlCheerioParserMiddleware {
   async process(context, next) {
-    if (!context.contentType.startsWith("text/html")) {
-      await next();
-      return;
-    }
-    const htmlString = typeof context.content === "string" ? context.content : Buffer.from(context.content).toString("utf-8");
     try {
       logger.debug(`Parsing HTML content with Cheerio from ${context.source}`);
-      const $ = cheerio.load(htmlString);
+      const $ = cheerio.load(context.content);
       context.dom = $;
       await next();
     } catch (error) {
@@ -403,17 +407,15 @@ function createJSDOM(html, options) {
 class HtmlLinkExtractorMiddleware {
   /**
    * Processes the context to extract links from the sanitized HTML body.
-   * @param context The current processing context.
+   * @param context The current middleware context.
    * @param next Function to call the next middleware.
    */
   async process(context, next) {
     const $ = context.dom;
     if (!$) {
-      if (context.contentType.startsWith("text/html")) {
-        logger.warn(
-          `Skipping ${this.constructor.name}: context.dom is missing for HTML content. Ensure HtmlCheerioParserMiddleware runs before this.`
-        );
-      }
+      logger.warn(
+        `Skipping ${this.constructor.name}: context.dom is missing. Ensure HtmlCheerioParserMiddleware runs before this.`
+      );
       await next();
       return;
     }
@@ -460,11 +462,9 @@ class HtmlMetadataExtractorMiddleware {
   async process(context, next) {
     const $ = context.dom;
     if (!$) {
-      if (context.contentType.startsWith("text/html")) {
-        logger.warn(
-          `Skipping ${this.constructor.name}: context.dom is missing for HTML content. Ensure HtmlCheerioParserMiddleware runs before this.`
-        );
-      }
+      logger.warn(
+        `Skipping ${this.constructor.name}: context.dom is missing. Ensure HtmlCheerioParserMiddleware runs before this.`
+      );
       await next();
       return;
     }
@@ -526,10 +526,6 @@ class HtmlPlaywrightMiddleware {
     }
   }
   async process(context, next) {
-    if (!context.contentType.startsWith("text/html")) {
-      await next();
-      return;
-    }
     const scrapeMode = context.options?.scrapeMode ?? ScrapeMode.Auto;
     const shouldRunPlaywright = scrapeMode === ScrapeMode.Playwright || scrapeMode === ScrapeMode.Auto;
     if (!shouldRunPlaywright) {
@@ -552,7 +548,7 @@ class HtmlPlaywrightMiddleware {
         if (route.request().url() === context.source) {
           return route.fulfill({
             status: 200,
-            contentType: context.contentType,
+            contentType: "text/html",
             body: context.content
           });
         }
@@ -655,11 +651,9 @@ class HtmlSanitizerMiddleware {
   async process(context, next) {
     const $ = context.dom;
     if (!$) {
-      if (context.contentType.startsWith("text/html")) {
-        logger.warn(
-          `Skipping ${this.constructor.name}: context.dom is missing. Ensure HtmlCheerioParserMiddleware runs before this.`
-        );
-      }
+      logger.warn(
+        `Skipping ${this.constructor.name}: context.dom is missing. Ensure HtmlCheerioParserMiddleware runs before this.`
+      );
       await next();
       return;
     }
@@ -769,11 +763,9 @@ ${text.replace(/^\n+|\n+$/g, "")}
   async process(context, next) {
     const $ = context.dom;
     if (!$) {
-      if (context.contentType.startsWith("text/html")) {
-        logger.warn(
-          `Skipping ${this.constructor.name}: context.dom is missing for HTML content. Ensure HtmlCheerioParserMiddleware ran correctly.`
-        );
-      }
+      logger.warn(
+        `Skipping ${this.constructor.name}: context.dom is missing. Ensure HtmlCheerioParserMiddleware ran correctly.`
+      );
       await next();
       return;
     }
@@ -785,10 +777,8 @@ ${text.replace(/^\n+|\n+$/g, "")}
         const warnMsg = `HTML to Markdown conversion resulted in empty content for ${context.source}.`;
         logger.warn(warnMsg);
         context.content = "";
-        context.contentType = "text/markdown";
       } else {
         context.content = markdown;
-        context.contentType = "text/markdown";
         logger.debug(`Successfully converted HTML to Markdown for ${context.source}`);
       }
     } catch (error) {
@@ -809,10 +799,8 @@ class MarkdownLinkExtractorMiddleware {
    * @param next Function to call the next middleware.
    */
   async process(context, next) {
-    if (context.contentType === "text/markdown") {
-      if (!Array.isArray(context.links)) {
-        context.links = [];
-      }
+    if (!Array.isArray(context.links)) {
+      context.links = [];
     }
     await next();
   }
@@ -824,31 +812,153 @@ class MarkdownMetadataExtractorMiddleware {
    * @param next Function to call the next middleware.
    */
   async process(context, next) {
-    if (context.contentType === "text/markdown" || context.contentType === "text/plain") {
-      try {
-        const textContent = typeof context.content === "string" ? context.content : Buffer.from(context.content).toString("utf-8");
-        if (typeof context.content !== "string") {
-          context.content = textContent;
-        }
-        let title = "Untitled";
-        if (context.contentType === "text/markdown") {
-          const match = textContent.match(/^#\s+(.*)$/m);
-          if (match?.[1]) {
-            title = match[1].trim();
-          }
-        }
-        context.metadata.title = title;
-      } catch (error) {
-        context.errors.push(
-          new Error(
-            `Failed to extract metadata from Markdown: ${error instanceof Error ? error.message : String(error)}`
-          )
-        );
+    try {
+      let title = "Untitled";
+      const match = context.content.match(/^#\s+(.*)$/m);
+      if (match?.[1]) {
+        title = match[1].trim();
       }
+      context.metadata.title = title;
+    } catch (error) {
+      context.errors.push(
+        new Error(
+          `Failed to extract metadata from Markdown: ${error instanceof Error ? error.message : String(error)}`
+        )
+      );
     }
     await next();
   }
 }
+function convertToString(content, charset) {
+  if (Buffer.isBuffer(content)) {
+    const decoder = new TextDecoder(charset || "utf-8");
+    return decoder.decode(content);
+  }
+  return content;
+}
+class BasePipeline {
+  /**
+   * Determines if this pipeline can process the given content.
+   * Must be implemented by derived classes.
+   */
+  canProcess(_rawContent) {
+    throw new Error("Method not implemented.");
+  }
+  /**
+   * Processes the raw content through the pipeline.
+   * Must be implemented by derived classes.
+   */
+  async process(_rawContent, _options, _fetcher) {
+    throw new Error("Method not implemented.");
+  }
+  /**
+   * Executes a middleware stack on the given context.
+   * This is a utility method used by derived pipeline classes.
+   *
+   * @param middleware - The middleware stack to execute
+   * @param context - The context to process
+   */
+  async executeMiddlewareStack(middleware, context) {
+    let index = -1;
+    const dispatch = async (i) => {
+      if (i <= index) throw new Error("next() called multiple times");
+      index = i;
+      const mw = middleware[i];
+      if (!mw) return;
+      await mw.process(context, dispatch.bind(null, i + 1));
+    };
+    try {
+      await dispatch(0);
+    } catch (error) {
+      context.errors.push(error instanceof Error ? error : new Error(String(error)));
+    }
+  }
+  /**
+   * Cleans up resources when the pipeline is no longer needed.
+   * Default implementation does nothing.
+   */
+  async close() {
+  }
+}
+class HtmlPipeline extends BasePipeline {
+  playwrightMiddleware;
+  standardMiddleware;
+  constructor() {
+    super();
+    this.playwrightMiddleware = new HtmlPlaywrightMiddleware();
+    this.standardMiddleware = [
+      new HtmlCheerioParserMiddleware(),
+      new HtmlMetadataExtractorMiddleware(),
+      new HtmlLinkExtractorMiddleware(),
+      new HtmlSanitizerMiddleware(),
+      new HtmlToMarkdownMiddleware()
+    ];
+  }
+  canProcess(rawContent) {
+    return MimeTypeUtils.isHtml(rawContent.mimeType);
+  }
+  async process(rawContent, options, fetcher) {
+    const contentString = convertToString(rawContent.content, rawContent.charset);
+    const context = {
+      content: contentString,
+      source: rawContent.source,
+      metadata: {},
+      links: [],
+      errors: [],
+      options,
+      fetcher
+    };
+    let middleware = [...this.standardMiddleware];
+    if (options.scrapeMode === "playwright" || options.scrapeMode === "auto") {
+      middleware = [this.playwrightMiddleware, ...middleware];
+    }
+    await this.executeMiddlewareStack(middleware, context);
+    return {
+      textContent: typeof context.content === "string" ? context.content : "",
+      metadata: context.metadata,
+      links: context.links,
+      errors: context.errors
+    };
+  }
+  async close() {
+    await this.playwrightMiddleware.closeBrowser();
+  }
+}
+class MarkdownPipeline extends BasePipeline {
+  middleware;
+  constructor() {
+    super();
+    this.middleware = [
+      new MarkdownMetadataExtractorMiddleware(),
+      new MarkdownLinkExtractorMiddleware()
+    ];
+  }
+  canProcess(rawContent) {
+    if (!rawContent.mimeType) return false;
+    return MimeTypeUtils.isMarkdown(rawContent.mimeType) || MimeTypeUtils.isText(rawContent.mimeType);
+  }
+  async process(rawContent, options, fetcher) {
+    const contentString = convertToString(rawContent.content, rawContent.charset);
+    const context = {
+      content: contentString,
+      source: rawContent.source,
+      metadata: {},
+      links: [],
+      errors: [],
+      options,
+      fetcher
+    };
+    await this.executeMiddlewareStack(this.middleware, context);
+    return {
+      textContent: typeof context.content === "string" ? context.content : "",
+      metadata: context.metadata,
+      links: context.links,
+      errors: context.errors
+    };
+  }
+  async close() {
+  }
+}
 class PipelineError extends Error {
   constructor(message, cause) {
     super(message);
@@ -976,12 +1086,15 @@ class BaseScraperStrategy {
 class WebScraperStrategy extends BaseScraperStrategy {
   httpFetcher = new HttpFetcher();
   shouldFollowLinkFn;
-  playwrightMiddleware;
-  // Add member
+  htmlPipeline;
+  markdownPipeline;
+  pipelines;
   constructor(options = {}) {
     super({ urlNormalizerOptions: options.urlNormalizerOptions });
     this.shouldFollowLinkFn = options.shouldFollowLink;
-    this.playwrightMiddleware = new HtmlPlaywrightMiddleware();
+    this.htmlPipeline = new HtmlPipeline();
+    this.markdownPipeline = new MarkdownPipeline();
+    this.pipelines = [this.htmlPipeline, this.markdownPipeline];
   }
   canHandle(url) {
     try {
@@ -1015,54 +1128,28 @@ class WebScraperStrategy extends BaseScraperStrategy {
         followRedirects: options.followRedirects
       };
       const rawContent = await this.httpFetcher.fetch(url, fetchOptions);
-      const initialContext = {
-        content: rawContent.content,
-        contentType: rawContent.mimeType,
-        source: rawContent.source,
-        // Use the final source URL after redirects
-        metadata: {},
-        links: [],
-        errors: [],
-        options,
-        fetcher: this.httpFetcher
-      };
-      let pipeline;
-      if (initialContext.contentType.startsWith("text/html")) {
-        const htmlPipelineSteps = [
-          this.playwrightMiddleware,
-          // Use the instance member
-          // TODO: Add HtmlJsExecutorMiddleware here if needed based on options
-          new HtmlCheerioParserMiddleware(),
-          // Always runs after content is finalized
-          new HtmlMetadataExtractorMiddleware(),
-          new HtmlLinkExtractorMiddleware(),
-          new HtmlSanitizerMiddleware(),
-          // Element remover
-          new HtmlToMarkdownMiddleware()
-        ];
-        pipeline = new ContentProcessingPipeline(htmlPipelineSteps);
-      } else if (initialContext.contentType === "text/markdown" || initialContext.contentType === "text/plain") {
-        pipeline = new ContentProcessingPipeline([
-          new MarkdownMetadataExtractorMiddleware(),
-          new MarkdownLinkExtractorMiddleware()
-          // Placeholder for now
-        ]);
-      } else {
+      let processed;
+      for (const pipeline of this.pipelines) {
+        if (pipeline.canProcess(rawContent)) {
+          processed = await pipeline.process(rawContent, options, this.httpFetcher);
+          break;
+        }
+      }
+      if (!processed) {
         logger.warn(
-          `Unsupported content type "${initialContext.contentType}" for URL ${url}. Skipping processing.`
+          `Unsupported content type "${rawContent.mimeType}" for URL ${url}. Skipping processing.`
         );
         return { document: void 0, links: [] };
       }
-      const finalContext = await pipeline.run(initialContext);
-      for (const err of finalContext.errors) {
+      for (const err of processed.errors) {
         logger.warn(`Processing error for ${url}: ${err.message}`);
       }
-      if (typeof finalContext.content !== "string" || !finalContext.content.trim()) {
+      if (!processed.textContent || !processed.textContent.trim()) {
         logger.warn(`No processable content found for ${url} after pipeline execution.`);
-        return { document: void 0, links: finalContext.links };
+        return { document: void 0, links: processed.links };
       }
       const baseUrl = new URL(options.url);
-      const filteredLinks = finalContext.links.filter((link) => {
+      const filteredLinks = processed.links.filter((link) => {
         try {
           const targetUrl = new URL(link);
           const scope = options.scope || "subpages";
@@ -1073,20 +1160,16 @@ class WebScraperStrategy extends BaseScraperStrategy {
       });
       return {
         document: {
-          content: finalContext.content,
-          // Final processed content (Markdown)
+          content: processed.textContent,
           metadata: {
-            url: finalContext.source,
-            // URL after redirects
-            // Ensure title is a string, default to "Untitled"
-            title: typeof finalContext.metadata.title === "string" ? finalContext.metadata.title : "Untitled",
+            url,
+            title: typeof processed.metadata.title === "string" ? processed.metadata.title : "Untitled",
             library: options.library,
-            version: options.version
-            // Add other metadata from context if needed
+            version: options.version,
+            ...processed.metadata
           }
         },
         links: filteredLinks
-        // Use the filtered links
       };
     } catch (error) {
       logger.error(`Failed processing page ${url}: ${error}`);
@@ -1101,7 +1184,8 @@ class WebScraperStrategy extends BaseScraperStrategy {
     try {
       await super.scrape(options, progressCallback, signal);
     } finally {
-      await this.playwrightMiddleware.closeBrowser();
+      await this.htmlPipeline.close();
+      await this.markdownPipeline.close();
     }
   }
 }
@@ -1153,6 +1237,15 @@ class GitHubScraperStrategy {
 }
 class LocalFileStrategy extends BaseScraperStrategy {
   fileFetcher = new FileFetcher();
+  htmlPipeline;
+  markdownPipeline;
+  pipelines;
+  constructor() {
+    super();
+    this.htmlPipeline = new HtmlPipeline();
+    this.markdownPipeline = new MarkdownPipeline();
+    this.pipelines = [this.htmlPipeline, this.markdownPipeline];
+  }
   canHandle(url) {
     return url.startsWith("file://");
   }
@@ -1167,62 +1260,41 @@ class LocalFileStrategy extends BaseScraperStrategy {
     }
     logger.info(`📄 Processing file ${this.pageCount}/${options.maxPages}: ${filePath}`);
     const rawContent = await this.fileFetcher.fetch(item.url);
-    const initialContext = {
-      content: rawContent.content,
-      contentType: rawContent.mimeType,
-      source: rawContent.source,
-      // file:// URL
-      metadata: {},
-      links: [],
-      // LocalFileStrategy doesn't extract links from file content itself
-      errors: [],
-      options
-      // Pass the full options object
-    };
-    let pipeline;
-    if (initialContext.contentType.startsWith("text/html")) {
-      pipeline = new ContentProcessingPipeline([
-        new HtmlCheerioParserMiddleware(),
-        new HtmlMetadataExtractorMiddleware(),
-        // No HtmlLinkExtractorMiddleware needed for local files
-        new HtmlSanitizerMiddleware(),
-        new HtmlToMarkdownMiddleware()
-      ]);
-    } else if (initialContext.contentType === "text/markdown" || initialContext.contentType === "text/plain" || // Treat plain text as markdown
-    initialContext.contentType.startsWith("text/")) {
-      pipeline = new ContentProcessingPipeline([
-        new MarkdownMetadataExtractorMiddleware()
-        // No MarkdownLinkExtractorMiddleware needed for local files
-      ]);
-    } else {
+    let processed;
+    for (const pipeline of this.pipelines) {
+      if (pipeline.canProcess(rawContent)) {
+        processed = await pipeline.process(rawContent, options, this.fileFetcher);
+        break;
+      }
+    }
+    if (!processed) {
       logger.warn(
-        `Unsupported content type "${initialContext.contentType}" for file ${filePath}. Skipping processing.`
+        `Unsupported content type "${rawContent.mimeType}" for file ${filePath}. Skipping processing.`
       );
       return { document: void 0, links: [] };
     }
-    const finalContext = await pipeline.run(initialContext);
-    for (const err of finalContext.errors) {
+    for (const err of processed.errors) {
       logger.warn(`Processing error for ${filePath}: ${err.message}`);
     }
-    const finalContentString = typeof finalContext.content === "string" ? finalContext.content : Buffer.from(finalContext.content).toString("utf-8");
     return {
       document: {
-        // Use the potentially empty string content
-        content: finalContentString,
+        content: typeof processed.textContent === "string" ? processed.textContent : "",
         metadata: {
-          url: finalContext.source,
-          // Use context source (file:// URL)
-          // Ensure title is a string, default to "Untitled"
-          title: typeof finalContext.metadata.title === "string" ? finalContext.metadata.title : "Untitled",
+          url: rawContent.source,
+          title: typeof processed.metadata.title === "string" ? processed.metadata.title : "Untitled",
           library: options.library,
           version: options.version
         }
       }
-      // No links returned from file content processing
     };
   }
   async scrape(options, progressCallback, signal) {
-    await super.scrape(options, progressCallback, signal);
+    try {
+      await super.scrape(options, progressCallback, signal);
+    } finally {
+      await this.htmlPipeline.close();
+      await this.markdownPipeline.close();
+    }
   }
 }
 class NpmScraperStrategy {
@@ -1456,7 +1528,9 @@ class PipelineManager {
     };
     this.jobMap.set(jobId, job);
     this.jobQueue.push(jobId);
-    logger.info(`📝 Job enqueued: ${jobId} for ${library}@${version}`);
+    logger.info(
+      `📝 Job enqueued: ${jobId} for ${library}${version ? `@${version}` : ""}`
+    );
     await this.callbacks.onJobStatusChange?.(job);
     if (this.isRunning) {
       this._processQueue();
@@ -2758,7 +2832,7 @@ class DocumentStore {
    */
   async initializeEmbeddings() {
     const modelSpec = process.env.DOCS_MCP_EMBEDDING_MODEL || "text-embedding-3-small";
-    const { createEmbeddingModel } = await import("./EmbeddingFactory-DZKXkqOe.js");
+    const { createEmbeddingModel } = await import("./EmbeddingFactory-Dz1hdJJe.js");
     this.embeddings = createEmbeddingModel(modelSpec);
     const testVector = await this.embeddings.embedQuery("test");
     this.modelDimension = testVector.length;
@@ -2873,7 +2947,12 @@ class DocumentStore {
 `;
         return `${header}${doc.pageContent}`;
       });
-      const rawEmbeddings = await this.embeddings.embedDocuments(texts);
+      const rawEmbeddings = [];
+      for (let i = 0; i < texts.length; i += EMBEDDING_BATCH_SIZE) {
+        const batchTexts = texts.slice(i, i + EMBEDDING_BATCH_SIZE);
+        const batchEmbeddings = await this.embeddings.embedDocuments(batchTexts);
+        rawEmbeddings.push(...batchEmbeddings);
+      }
       const paddedEmbeddings = rawEmbeddings.map((vector) => this.padVector(vector));
       const transaction = this.db.transaction((docs) => {
         for (let i = 0; i < docs.length; i++) {
@@ -3372,12 +3451,11 @@ class DocumentManagementService {
   }
 }
 export {
-  ContentProcessingPipeline as C,
   DocumentManagementService as D,
   FileFetcher as F,
   HttpFetcher as H,
   LibraryNotFoundError as L,
-  MarkdownMetadataExtractorMiddleware as M,
+  MarkdownPipeline as M,
   PipelineJobStatus as P,
   SearchTool as S,
   ToolError as T,
@@ -3392,18 +3470,14 @@ export {
   DEFAULT_HTTP_PORT as h,
   DEFAULT_MAX_CONCURRENCY as i,
   ScrapeMode as j,
-  HtmlPlaywrightMiddleware as k,
+  HtmlPipeline as k,
   logger as l,
-  HtmlCheerioParserMiddleware as m,
-  HtmlMetadataExtractorMiddleware as n,
-  HtmlSanitizerMiddleware as o,
-  HtmlToMarkdownMiddleware as p,
-  ScraperError as q,
-  createJSDOM as r,
-  setLogLevel as s,
-  getProjectRoot as t,
-  DEFAULT_WEB_PORT as u,
-  DimensionError as v,
-  VECTOR_DIMENSION as w
+  ScraperError as m,
+  createJSDOM as n,
+  getProjectRoot as o,
+  DEFAULT_WEB_PORT as p,
+  DimensionError as q,
+  VECTOR_DIMENSION as r,
+  setLogLevel as s
 };
-//# sourceMappingURL=DocumentManagementService-BupnR1eC.js.map
+//# sourceMappingURL=DocumentManagementService-BZ_ZZgPI.js.map