@absolutejs/rag 0.15.0 → 0.15.1

package/CHANGELOG.md

@@ -6,6 +6,12 @@ This file is generated by `absolute-changelog` from the entries in
 `changelog/`. Edit an entry, not this file — and add new ones under
 `changelog/unreleased/`.
 
+## 0.15.1 — 2026-09-20
+
+### Fixed
+
+- **Preserve all sections of public websites instead of extracting one article fragment, and retain labeled image evidence in source-bound documents.** (`readRAGWebpage`, `readRAGWebsite`)
+
 ## 0.15.0 — 2026-09-18
 
 ### Added

package/changelog.json

@@ -2,6 +2,20 @@
 	"contract": 1,
 	"name": "@absolutejs/rag",
 	"releases": [
+		{
+			"changes": [
+				{
+					"kind": "fixed",
+					"summary": "Preserve all sections of public websites instead of extracting one article fragment, and retain labeled image evidence in source-bound documents.",
+					"symbols": [
+						"readRAGWebpage",
+						"readRAGWebsite"
+					]
+				}
+			],
+			"date": "2026-09-20",
+			"version": "0.15.1"
+		},
 		{
 			"changes": [
 				{

package/dist/src/web/text.d.ts

@@ -0,0 +1,2 @@
+/** Website research needs the whole page, not the highest-scoring article fragment. */
+export declare const websiteHtmlText: (html: string) => string;

package/dist/web/index.js

@@ -11256,6 +11256,34 @@ var hasIncompleteAppContent = (html) => {
   return text.length < 100 || /^(?:loading|please wait)(?:[.\s]|$)/iu.test(text);
 };
 
+// src/web/text.ts
+var websiteHtmlText = (html) => {
+  const { document } = parseHTML(html);
+  document.querySelectorAll("script,style,template,noscript,svg,[hidden],[aria-hidden=true]").forEach((element) => element.remove());
+  const root = document.body?.textContent?.trim() ? document.body : document.documentElement;
+  const parts = [];
+  const blocks = /^(?:ADDRESS|ARTICLE|ASIDE|BLOCKQUOTE|BR|DIV|DL|DT|DD|FIGCAPTION|FIGURE|FOOTER|H[1-6]|HEADER|HR|LI|MAIN|NAV|OL|P|SECTION|TABLE|TD|TH|TR|UL)$/u;
+  const walk = (node2) => {
+    if (node2.nodeType === 3) {
+      parts.push((node2.textContent ?? "").replace(/\s+/gu, " "));
+      return;
+    }
+    const block = blocks.test(node2.nodeName);
+    if (block)
+      parts.push(`
+`);
+    node2.childNodes.forEach(walk);
+    if (block)
+      parts.push(`
+`);
+  };
+  if (root)
+    walk(root);
+  return parts.join("").split(`
+`).map((line) => line.trim()).filter(Boolean).join(`
+`);
+};
+
 // src/ingestion/ingestion.ts
 import { readdir, readFile } from "fs/promises";
 import { basename, extname, join, relative, resolve } from "path";
@@ -17803,13 +17831,22 @@ var readRAGWebsite = async (options) => {
     const page = pages.find((item) => item.finalUrl === source.url && item.text);
     if (!page || remainingDocumentChars <= 0)
       return [];
-    const text = page.text.slice(0, remainingDocumentChars);
+    const imageLabels = page.evidence.images.filter((image) => image.label).map((image) => `${image.context ? `${image.context}: ` : ""}${image.label}`);
+    const fullText = [
+      page.text,
+      ...imageLabels.length ? [
+        "Image labels supplied by the website (not independently verified customer relationships):",
+        ...imageLabels
+      ] : []
+    ].join(`
+`);
+    const text = fullText.slice(0, remainingDocumentChars);
     remainingDocumentChars -= text.length;
     return [
       {
         sourceId: source.id,
         text,
-        truncated: page.truncated || text.length < page.text.length
+        truncated: page.truncated || text.length < fullText.length
       }
     ];
   });
@@ -17856,7 +17893,7 @@ ${page.text}`).join(`
 // src/web/index.ts
 var messageFor = (error) => error instanceof Error ? error.message : "Website retrieval failed.";
 var codeFor = (error) => error instanceof WebReadError ? error.code : error instanceof Error && /^(?:AbortError|TimeoutError)$/u.test(error.name) ? "timeout" : "fetch_failed";
-var cleanHtml = (html) => prepareRAGDocument({ text: html, format: "html" }).normalizedText;
+var cleanHtml = websiteHtmlText;
 var titleOf = (html) => {
   const title = /<title\b[^>]*>([\s\S]*?)<\/title>/iu.exec(html)?.[1];
   return title ? prepareRAGDocument({ text: title, format: "html" }).normalizedText : null;
@@ -18025,5 +18062,5 @@ export {
   validatePublicWebUrl
 };
 
-//# debugId=3F4B779525714ED064756E2164756E21
+//# debugId=64305FA75E0E709664756E2164756E21
 //# sourceMappingURL=index.js.map