@mastra/rag 2.0.0-beta.2 → 2.0.0-beta.3

package/dist/index.js

@@ -266,7 +266,7 @@ Provide keywords in the following comma-separated format: 'KEYWORDS: <keywords>'
 var defaultQuestionExtractPrompt = new PromptTemplate({
   templateVars: ["numQuestions", "context"],
   template: `(
-  "Given the contextual informations below, generate {numQuestions} questions this context can provides specific answers to which are unlikely to be found else where. Higher-level summaries of surrounding context may be provided as well. "
+  "Given the contextual information below, generate {numQuestions} questions this context can provide specific answers to which are unlikely to be found elsewhere. Higher-level summaries of surrounding context may be provided as well. "
   "Try using these summaries to generate better questions that this context can answer."
   "---------------------"
   "{context}"
@@ -4993,23 +4993,47 @@ var HTMLSectionTransformer = class {
     }
     return "/" + parts.join("/");
   }
+  getTextContent(element) {
+    if (!element) return "";
+    if (!element.tagName) {
+      return element.text || "";
+    }
+    let content = element.text || "";
+    if (element.childNodes) {
+      for (const child of element.childNodes) {
+        const childText = this.getTextContent(child);
+        if (childText) {
+          content += " " + childText;
+        }
+      }
+    }
+    return content.trim();
+  }
   splitHtmlByHeaders(htmlDoc) {
     const sections = [];
     const root = parse(htmlDoc);
     const headers = Object.keys(this.headersToSplitOn);
     const headerElements = root.querySelectorAll(headers.join(","));
-    headerElements.forEach((headerElement, index) => {
+    headerElements.forEach((headerElement) => {
       const header = headerElement.text?.trim() || "";
       const tagName = headerElement.tagName;
       const xpath = this.getXPath(headerElement);
       let content = "";
-      let currentElement = headerElement.nextElementSibling;
-      const nextHeader = headerElements[index + 1];
-      while (currentElement && (!nextHeader || currentElement !== nextHeader)) {
-        if (currentElement.text) {
-          content += currentElement.text.trim() + " ";
+      const parentNode = headerElement.parentNode;
+      if (parentNode && parentNode.childNodes) {
+        let foundHeader = false;
+        for (const node of parentNode.childNodes) {
+          if (node === headerElement) {
+            foundHeader = true;
+            continue;
+          }
+          if (foundHeader && node.tagName && headers.includes(node.tagName.toLowerCase())) {
+            break;
+          }
+          if (foundHeader) {
+            content += this.getTextContent(node) + " ";
+          }
         }
-        currentElement = currentElement.nextElementSibling;
       }
       content = content.trim();
       sections.push({
@@ -6386,13 +6410,33 @@ var MDocument = class _MDocument {
   async chunkHTML(options) {
     if (options?.headers?.length) {
       const rt = new HTMLHeaderTransformer(options);
-      const textSplit = rt.transformDocuments(this.chunks);
+      let textSplit = rt.transformDocuments(this.chunks);
+      if (options?.maxSize) {
+        const textSplitter = new RecursiveCharacterTransformer({
+          maxSize: options.maxSize,
+          overlap: options.overlap,
+          keepSeparator: options.keepSeparator,
+          addStartIndex: options.addStartIndex,
+          stripWhitespace: options.stripWhitespace
+        });
+        textSplit = textSplitter.splitDocuments(textSplit);
+      }
       this.chunks = textSplit;
       return;
     }
     if (options?.sections?.length) {
       const rt = new HTMLSectionTransformer(options);
-      const textSplit = rt.transformDocuments(this.chunks);
+      let textSplit = rt.transformDocuments(this.chunks);
+      if (options?.maxSize) {
+        const textSplitter = new RecursiveCharacterTransformer({
+          maxSize: options.maxSize,
+          overlap: options.overlap,
+          keepSeparator: options.keepSeparator,
+          addStartIndex: options.addStartIndex,
+          stripWhitespace: options.stripWhitespace
+        });
+        textSplit = textSplitter.splitDocuments(textSplit);
+      }
       this.chunks = textSplit;
       return;
     }