fix: derive word_count from the content actually resolved

README.md

@@ -271,6 +271,12 @@ Both templates support `id` autocompletion: as you type, the server searches Boo
 4. In the **API Tokens** section, create a new token
 5. Copy the Token ID and Token Secret
 
+> **Also grant the role “Export Content”** if you read pages written in the
+> WYSIWYG editor. BookStack returns an empty `markdown` body for those pages, and
+> `get_page` recovers it from the server-side HTML→markdown export endpoint. Without
+> the permission that fallback fails and the page reads as empty. The default Viewer
+> role does not include it.
+
 ## Security
 
 - Write operations are **disabled by default**

src/bookstack-client.ts

@@ -1,6 +1,7 @@
 import axios, { AxiosInstance, AxiosError, InternalAxiosRequestConfig, AxiosAdapter } from 'axios';
 import https from 'https';
 import { Semaphore } from './util/semaphore.js';
+import { countWords } from './util/word-count.js';
 
 const MAX_RETRIES_429 = 5;
 
@@ -402,7 +403,7 @@ export class BookStackClient {
     return {
       ...pageMeta,
       url,
-      word_count: page.text ? page.text.split(' ').length : 0,
+      word_count: countWords(page),
       content_format: format,
       content_total_chars: totalChars,
       content_offset: offset,

src/util/word-count.test.ts

@@ -0,0 +1,38 @@
+import test from 'node:test';
+import assert from 'node:assert/strict';
+import { countWords } from './word-count.js';
+
+test('counts words in page.text when BookStack populates it', () => {
+  assert.equal(countWords({ text: 'one two three' }), 3);
+});
+
+test('splits on any whitespace, not just a single space', () => {
+  assert.equal(countWords({ text: 'one\ntwo\tthree  four\r\nfive' }), 5);
+});
+
+test('ignores leading, trailing and repeated whitespace', () => {
+  assert.equal(countWords({ text: '  one   two  ' }), 2);
+});
+
+test('falls back to markdown when text is empty (WYSIWYG pages)', () => {
+  assert.equal(countWords({ text: '', markdown: '# Title\n\nBody text here' }), 5);
+});
+
+test('falls back to html when neither text nor markdown is present', () => {
+  assert.equal(countWords({ html: '<p>Hello <strong>there</strong> world</p>' }), 3);
+});
+
+test('does not count markup or script/style bodies as words', () => {
+  const html = '<style>.a{color:red}</style><script>var x = 1;</script><p>only these three</p>';
+  assert.equal(countWords({ html }), 3);
+});
+
+test('treats &nbsp; as a separator', () => {
+  assert.equal(countWords({ html: '<p>one&nbsp;two</p>' }), 2);
+});
+
+test('returns 0 for a page with no content at all', () => {
+  assert.equal(countWords({}), 0);
+  assert.equal(countWords({ text: '', markdown: '', html: '' }), 0);
+  assert.equal(countWords({ text: null, markdown: null, html: null }), 0);
+});

src/util/word-count.ts

@@ -0,0 +1,23 @@
+// BookStack only populates page.text for markdown-authored pages. For pages
+// written in the WYSIWYG editor it comes back empty, so a word count taken
+// from page.text alone reports 0 for pages that plainly have content (#10).
+// Fall back to whichever body the response actually carries.
+
+function stripHtml(html: string): string {
+  return html
+    .replace(/<(script|style)\b[^>]*>[\s\S]*?<\/\1>/gi, ' ')
+    .replace(/<[^>]+>/g, ' ')
+    .replace(/&nbsp;/gi, ' ');
+}
+
+export function countWords(page: {
+  text?: string | null;
+  markdown?: string | null;
+  html?: string | null;
+}): number {
+  const source =
+    page.text || page.markdown || (page.html ? stripHtml(page.html) : '');
+  // Split on any whitespace, not a single space: a page whose lines are
+  // newline-separated counts as one word under /' '/.
+  return source.split(/\s+/).filter(Boolean).length;
+}