create-nextblock 0.16.1 → 0.16.2

package/templates/nextblock-template/lib/seo/page-document.test.ts

@@ -1,350 +1,350 @@
-import { describe, expect, it } from 'vitest';
-import { auditSeo, emptySeoDocument, tokenizeWords } from '@nextblock-cms/utils/seo';
-import { buildPageSeoDocument } from './page-document';
-
-/**
- * The walker is the piece that decides what the page-level audit is allowed to
- * see, so these tests are written against the block shapes the CMS actually
- * stores — `{ block_type, content }` rows, with a section's children nested in
- * `content.column_blocks[column][index]` rather than sitting in the row list.
- *
- * Two things are asserted over and over on purpose. Content must be *found*,
- * because a walker that quietly misses a container under-reports the page and
- * hands the author a score for copy that was never read; and malformed content
- * must be *survived*, because this runs on every keystroke over half-typed
- * editor state where any field can be null or the wrong type.
- */
-
-/** A rich-text block as the editor stores it when the content is HTML. */
-function textBlock(html: string): unknown {
-  return { block_type: 'text', content: { html_content: html } };
-}
-
-/** A standalone heading block, the shape a per-block audit could never see. */
-function headingBlock(level: number, text: string): unknown {
-  return { block_type: 'heading', content: { level, text_content: text } };
-}
-
-/** A section, whose children are plain objects inside its columns. */
-function sectionBlock(columns: unknown[][]): unknown {
-  return { block_type: 'section', content: { column_blocks: columns } };
-}
-
-describe('buildPageSeoDocument on an empty or unusable input', () => {
-  it('returns an empty document rather than throwing', () => {
-    const empty = emptySeoDocument();
-
-    expect(buildPageSeoDocument([])).toEqual(empty);
-    expect(buildPageSeoDocument(null)).toEqual(empty);
-    expect(buildPageSeoDocument(undefined)).toEqual(empty);
-    expect(buildPageSeoDocument('not blocks')).toEqual(empty);
-    expect(buildPageSeoDocument({ block_type: 'text' })).toEqual(empty);
-  });
-});
-
-describe('buildPageSeoDocument across sibling blocks', () => {
-  const document = buildPageSeoDocument([
-    headingBlock(1, 'Fresh Cat Food'),
-    textBlock('<p>We ship every order fast.</p>'),
-    headingBlock(2, 'Why It Works'),
-  ]);
-
-  it('merges every block into a single document with the headings in order', () => {
-    expect(document.headings).toEqual([
-      { level: 1, order: 0, text: 'Fresh Cat Food' },
-      { level: 2, order: 1, text: 'Why It Works' },
-    ]);
-  });
-
-  it('renumbers `order` sequentially across the whole page', () => {
-    // Each block starts its own numbering, so the page-level order can only
-    // come from the walker. It is the number the audit cites when it tells an
-    // author where their H1 sits, so an off-by-one here points at the wrong
-    // heading.
-    const withInlineHeadings = buildPageSeoDocument([
-      textBlock('<h2>Inline one</h2><h3>Inline two</h3>'),
-      headingBlock(2, 'Standalone three'),
-    ]);
-
-    expect(withInlineHeadings.headings.map((heading) => heading.order)).toEqual([0, 1, 2]);
-    expect(withInlineHeadings.headings[2]?.text).toBe('Standalone three');
-  });
-
-  it('keeps the text of adjacent blocks from gluing into one word', () => {
-    // "…our coffee" followed by "beans are…" must not tokenise as "coffeebeans":
-    // that single fabricated word would corrupt the word count, the keyphrase
-    // density and the readability sample in one go.
-    const glued = buildPageSeoDocument([
-      textBlock('<p>We roast our coffee</p>'),
-      textBlock('<p>beans are fresh</p>'),
-    ]);
-
-    expect(glued.words).toEqual(['we', 'roast', 'our', 'coffee', 'beans', 'are', 'fresh']);
-    expect(glued.text).toBe('We roast our coffee beans are fresh');
-  });
-
-  it('keeps `words` in step with `text`', () => {
-    // Every downstream metric reads one or the other and assumes they describe
-    // the same prose: readability tokenises `text`, the density divides by
-    // `words.length`.
-    expect(document.words).toEqual(tokenizeWords(document.text));
-  });
-
-  it('counts heading copy as page text', () => {
-    expect(document.text).toContain('Fresh Cat Food');
-    expect(document.words).toContain('fresh');
-  });
-});
-
-describe('buildPageSeoDocument on nested sections', () => {
-  it('finds blocks nested inside a section column', () => {
-    const document = buildPageSeoDocument([
-      sectionBlock([[headingBlock(1, 'Nested Title'), textBlock('<p>Nested copy.</p>')]]),
-    ]);
-
-    expect(document.headings).toEqual([{ level: 1, order: 0, text: 'Nested Title' }]);
-    expect(document.text).toBe('Nested Title Nested copy.');
-  });
-
-  it('reads column 0 before column 1', () => {
-    const document = buildPageSeoDocument([
-      sectionBlock([[headingBlock(2, 'Left')], [headingBlock(3, 'Right')]]),
-    ]);
-
-    expect(document.headings.map((heading) => heading.text)).toEqual(['Left', 'Right']);
-    expect(document.text).toBe('Left Right');
-  });
-
-  it('follows a section nested inside another section', () => {
-    const document = buildPageSeoDocument([
-      sectionBlock([[sectionBlock([[textBlock('<p>Two levels down.</p>')]])]]),
-    ]);
-
-    expect(document.text).toBe('Two levels down.');
-  });
-
-  it('reads a slider section from its slides instead of its columns', () => {
-    // A section in slider mode renders its slides and ignores whatever is left
-    // in `column_blocks`, so counting both would credit the page with copy no
-    // visitor can reach.
-    const document = buildPageSeoDocument([
-      {
-        block_type: 'section',
-        content: {
-          column_blocks: [[textBlock('<p>Stale hidden copy.</p>')]],
-          slider: true,
-          slides: [{ column_blocks: [[textBlock('<p>Visible slide copy.</p>')]] }],
-        },
-      },
-    ]);
-
-    expect(document.text).toBe('Visible slide copy.');
-  });
-
-  it('still reads a legacy hero block, which stores its children in slides', () => {
-    const document = buildPageSeoDocument([
-      {
-        block_type: 'hero',
-        content: { slides: [{ column_blocks: [[headingBlock(1, 'Legacy Hero')]] }] },
-      },
-    ]);
-
-    expect(document.headings).toEqual([{ level: 1, order: 0, text: 'Legacy Hero' }]);
-  });
-});
-
-describe('buildPageSeoDocument on rich-text blocks', () => {
-  it('reads an html_content that holds a JSON-stringified Tiptap document', () => {
-    // Both storage formats are live in the same column; the app tells them apart
-    // by a leading brace, and so must anything that reads the column.
-    const tiptap = JSON.stringify({
-      content: [
-        { attrs: { level: 2 }, content: [{ text: 'Tiptap Heading', type: 'text' }], type: 'heading' },
-        { content: [{ text: 'Tiptap paragraph copy.', type: 'text' }], type: 'paragraph' },
-      ],
-      type: 'doc',
-    });
-
-    const document = buildPageSeoDocument([textBlock(tiptap)]);
-
-    expect(document.headings).toEqual([{ level: 2, order: 0, text: 'Tiptap Heading' }]);
-    expect(document.text).toBe('Tiptap Heading Tiptap paragraph copy.');
-  });
-
-  it('collects images and links embedded in prose', () => {
-    const document = buildPageSeoDocument([
-      textBlock('<p><img src="/inline.png" alt=""><a href="https://example.com">Away</a></p>'),
-    ]);
-
-    expect(document.images).toEqual([{ alt: '', src: '/inline.png' }]);
-    expect(document.links).toEqual([
-      { external: true, href: 'https://example.com', text: 'Away' },
-    ]);
-  });
-});
-
-describe('buildPageSeoDocument on image, button and video blocks', () => {
-  it('reports an image block through the same alt-text channel as inline images', () => {
-    const document = buildPageSeoDocument([
-      { block_type: 'image', content: { alt_text: 'A bowl of cat food', object_key: 'cat.png' } },
-      { block_type: 'image', content: { alt_text: '', external_url: 'https://cdn.test/b.png' } },
-    ]);
-
-    expect(document.images).toEqual([
-      { alt: 'A bowl of cat food', src: 'cat.png' },
-      { alt: '', src: 'https://cdn.test/b.png' },
-    ]);
-  });
-
-  it('prefers the external URL, which is what the renderer draws', () => {
-    const document = buildPageSeoDocument([
-      {
-        block_type: 'image',
-        content: { alt_text: 'Stock', external_url: 'https://cdn.test/a.png', object_key: 'a.png' },
-      },
-    ]);
-
-    expect(document.images[0]?.src).toBe('https://cdn.test/a.png');
-  });
-
-  it('ignores an image block with no source at all', () => {
-    // That block renders a "media not selected" placeholder, so flagging its
-    // missing alt text would be a complaint about an image nobody can see.
-    const document = buildPageSeoDocument([{ block_type: 'image', content: { alt_text: '' } }]);
-
-    expect(document.images).toEqual([]);
-  });
-
-  it('includes an image caption in the page text', () => {
-    const document = buildPageSeoDocument([
-      {
-        block_type: 'image',
-        content: { alt_text: 'Beans', caption: 'Beans drying in the sun.', object_key: 'b.png' },
-      },
-    ]);
-
-    expect(document.text).toBe('Beans drying in the sun.');
-  });
-
-  it('records a button as a link and its label as text', () => {
-    const document = buildPageSeoDocument([
-      { block_type: 'button', content: { text: 'Shop now', url: '/shop' } },
-    ]);
-
-    expect(document.links).toEqual([{ external: false, href: '/shop', text: 'Shop now' }]);
-    expect(document.text).toBe('Shop now');
-  });
-
-  it('keeps a button label even when no URL has been set yet', () => {
-    const document = buildPageSeoDocument([{ block_type: 'button', content: { text: 'Shop now' } }]);
-
-    expect(document.links).toEqual([]);
-    expect(document.text).toBe('Shop now');
-  });
-
-  it('takes a video block title as text when there is one', () => {
-    const document = buildPageSeoDocument([
-      { block_type: 'video_embed', content: { title: 'How we roast', url: 'https://v.test/1' } },
-      { block_type: 'video_embed', content: { url: 'https://v.test/2' } },
-    ]);
-
-    expect(document.text).toBe('How we roast');
-  });
-
-  it('takes a testimonial quote and its attribution', () => {
-    const document = buildPageSeoDocument([
-      {
-        block_type: 'testimonial',
-        content: { author_name: 'Jane Doe', author_title: 'CEO', quote: 'It changed my morning.' },
-      },
-    ]);
-
-    expect(document.text).toBe('It changed my morning. Jane Doe CEO');
-  });
-});
-
-describe('buildPageSeoDocument on content it cannot read', () => {
-  it('skips unknown and custom block types instead of guessing at their fields', () => {
-    const document = buildPageSeoDocument([
-      { block_type: 'pricing-table-v2', content: { internal_label: 'Do not index me', rows: 4 } },
-      { block_type: 'posts_grid', content: { columns: 3, postsPerPage: 6 } },
-      headingBlock(2, 'Real Heading'),
-    ]);
-
-    expect(document.headings).toEqual([{ level: 2, order: 0, text: 'Real Heading' }]);
-    expect(document.text).toBe('Real Heading');
-  });
-
-  it('survives every field being the wrong type or absent', () => {
-    const malformed: unknown[] = [
-      null,
-      undefined,
-      42,
-      'a string where a row should be',
-      [],
-      { block_type: 'text' },
-      { block_type: 'text', content: null },
-      { block_type: 'text', content: { html_content: 12 } },
-      { block_type: 'heading', content: {} },
-      { block_type: 'image', content: { alt_text: null, object_key: null } },
-      { block_type: 'button', content: { text: null, url: null } },
-      { block_type: 'section', content: { column_blocks: 'not an array' } },
-      { block_type: 'section', content: { column_blocks: [null, [null, 7]] } },
-      { block_type: 'section', content: {} },
-      { content: { html_content: '<p>No block type.</p>' } },
-    ];
-
-    expect(() => buildPageSeoDocument(malformed)).not.toThrow();
-
-    const document = buildPageSeoDocument(malformed);
-
-    // The one heading with no fields at all still surfaces, because an empty
-    // heading is a real defect the audit reports rather than a row to discard.
-    expect(document.headings).toEqual([{ level: 2, order: 0, text: '' }]);
-    expect(document.text).toBe('');
-    expect(document.words).toEqual([]);
-  });
-
-  it('stops descending once the nesting is deeper than any real page', () => {
-    // Guards against a corrupt row that refers back into itself: the walk is
-    // bounded, so the editor degrades to missing copy rather than to a blown
-    // stack while someone is typing.
-    let nested: unknown = textBlock('<p>Deepest copy.</p>');
-    for (let level = 0; level < 40; level += 1) {
-      nested = sectionBlock([[nested]]);
-    }
-
-    expect(() => buildPageSeoDocument([nested])).not.toThrow();
-    expect(buildPageSeoDocument([nested]).text).toBe('');
-  });
-});
-
-describe('buildPageSeoDocument feeding the page-level audit', () => {
-  it('no longer reports a missing H1 when the H1 lives in its own block', () => {
-    // This is the bug that prompted the page-level audit: the paragraph block
-    // was graded on its own and reported "this page has no H1" while the H1 sat
-    // one row above it, in a block the auditor was never shown.
-    const blocks = [
-      headingBlock(1, 'Fresh Cat Food'),
-      sectionBlock([[textBlock('<p>We ship every order fast and pack it with care.</p>')]]),
-    ];
-
-    const paragraphOnly = auditSeo({ document: buildPageSeoDocument([blocks[1]]) });
-    const wholePage = auditSeo({ document: buildPageSeoDocument(blocks) });
-
-    expect(paragraphOnly.issues.map((issue) => issue.id)).toContain('headings-missing-h1');
-    expect(wholePage.issues.map((issue) => issue.id)).not.toContain('headings-missing-h1');
-    expect(wholePage.headings).toEqual([{ level: 1, order: 0, text: 'Fresh Cat Food' }]);
-  });
-
-  it('counts the words of the whole page rather than of one paragraph', () => {
-    const sentence = '<p>We ship every order fast and we pack it with care for your pet today.</p>';
-    const blocks = new Array(30).fill(null).map(() => textBlock(sentence));
-
-    const oneBlock = auditSeo({ document: buildPageSeoDocument([blocks[0]]) });
-    const wholePage = auditSeo({ document: buildPageSeoDocument(blocks) });
-
-    expect(oneBlock.issues.map((issue) => issue.id)).toContain('content-thin');
-    expect(wholePage.issues.map((issue) => issue.id)).not.toContain('content-thin');
-  });
-});
+import { describe, expect, it } from 'vitest';
+import { auditSeo, emptySeoDocument, tokenizeWords } from '@nextblock-cms/utils/seo';
+import { buildPageSeoDocument } from './page-document';
+
+/**
+ * The walker is the piece that decides what the page-level audit is allowed to
+ * see, so these tests are written against the block shapes the CMS actually
+ * stores — `{ block_type, content }` rows, with a section's children nested in
+ * `content.column_blocks[column][index]` rather than sitting in the row list.
+ *
+ * Two things are asserted over and over on purpose. Content must be *found*,
+ * because a walker that quietly misses a container under-reports the page and
+ * hands the author a score for copy that was never read; and malformed content
+ * must be *survived*, because this runs on every keystroke over half-typed
+ * editor state where any field can be null or the wrong type.
+ */
+
+/** A rich-text block as the editor stores it when the content is HTML. */
+function textBlock(html: string): unknown {
+  return { block_type: 'text', content: { html_content: html } };
+}
+
+/** A standalone heading block, the shape a per-block audit could never see. */
+function headingBlock(level: number, text: string): unknown {
+  return { block_type: 'heading', content: { level, text_content: text } };
+}
+
+/** A section, whose children are plain objects inside its columns. */
+function sectionBlock(columns: unknown[][]): unknown {
+  return { block_type: 'section', content: { column_blocks: columns } };
+}
+
+describe('buildPageSeoDocument on an empty or unusable input', () => {
+  it('returns an empty document rather than throwing', () => {
+    const empty = emptySeoDocument();
+
+    expect(buildPageSeoDocument([])).toEqual(empty);
+    expect(buildPageSeoDocument(null)).toEqual(empty);
+    expect(buildPageSeoDocument(undefined)).toEqual(empty);
+    expect(buildPageSeoDocument('not blocks')).toEqual(empty);
+    expect(buildPageSeoDocument({ block_type: 'text' })).toEqual(empty);
+  });
+});
+
+describe('buildPageSeoDocument across sibling blocks', () => {
+  const document = buildPageSeoDocument([
+    headingBlock(1, 'Fresh Cat Food'),
+    textBlock('<p>We ship every order fast.</p>'),
+    headingBlock(2, 'Why It Works'),
+  ]);
+
+  it('merges every block into a single document with the headings in order', () => {
+    expect(document.headings).toEqual([
+      { level: 1, order: 0, text: 'Fresh Cat Food' },
+      { level: 2, order: 1, text: 'Why It Works' },
+    ]);
+  });
+
+  it('renumbers `order` sequentially across the whole page', () => {
+    // Each block starts its own numbering, so the page-level order can only
+    // come from the walker. It is the number the audit cites when it tells an
+    // author where their H1 sits, so an off-by-one here points at the wrong
+    // heading.
+    const withInlineHeadings = buildPageSeoDocument([
+      textBlock('<h2>Inline one</h2><h3>Inline two</h3>'),
+      headingBlock(2, 'Standalone three'),
+    ]);
+
+    expect(withInlineHeadings.headings.map((heading) => heading.order)).toEqual([0, 1, 2]);
+    expect(withInlineHeadings.headings[2]?.text).toBe('Standalone three');
+  });
+
+  it('keeps the text of adjacent blocks from gluing into one word', () => {
+    // "…our coffee" followed by "beans are…" must not tokenise as "coffeebeans":
+    // that single fabricated word would corrupt the word count, the keyphrase
+    // density and the readability sample in one go.
+    const glued = buildPageSeoDocument([
+      textBlock('<p>We roast our coffee</p>'),
+      textBlock('<p>beans are fresh</p>'),
+    ]);
+
+    expect(glued.words).toEqual(['we', 'roast', 'our', 'coffee', 'beans', 'are', 'fresh']);
+    expect(glued.text).toBe('We roast our coffee beans are fresh');
+  });
+
+  it('keeps `words` in step with `text`', () => {
+    // Every downstream metric reads one or the other and assumes they describe
+    // the same prose: readability tokenises `text`, the density divides by
+    // `words.length`.
+    expect(document.words).toEqual(tokenizeWords(document.text));
+  });
+
+  it('counts heading copy as page text', () => {
+    expect(document.text).toContain('Fresh Cat Food');
+    expect(document.words).toContain('fresh');
+  });
+});
+
+describe('buildPageSeoDocument on nested sections', () => {
+  it('finds blocks nested inside a section column', () => {
+    const document = buildPageSeoDocument([
+      sectionBlock([[headingBlock(1, 'Nested Title'), textBlock('<p>Nested copy.</p>')]]),
+    ]);
+
+    expect(document.headings).toEqual([{ level: 1, order: 0, text: 'Nested Title' }]);
+    expect(document.text).toBe('Nested Title Nested copy.');
+  });
+
+  it('reads column 0 before column 1', () => {
+    const document = buildPageSeoDocument([
+      sectionBlock([[headingBlock(2, 'Left')], [headingBlock(3, 'Right')]]),
+    ]);
+
+    expect(document.headings.map((heading) => heading.text)).toEqual(['Left', 'Right']);
+    expect(document.text).toBe('Left Right');
+  });
+
+  it('follows a section nested inside another section', () => {
+    const document = buildPageSeoDocument([
+      sectionBlock([[sectionBlock([[textBlock('<p>Two levels down.</p>')]])]]),
+    ]);
+
+    expect(document.text).toBe('Two levels down.');
+  });
+
+  it('reads a slider section from its slides instead of its columns', () => {
+    // A section in slider mode renders its slides and ignores whatever is left
+    // in `column_blocks`, so counting both would credit the page with copy no
+    // visitor can reach.
+    const document = buildPageSeoDocument([
+      {
+        block_type: 'section',
+        content: {
+          column_blocks: [[textBlock('<p>Stale hidden copy.</p>')]],
+          slider: true,
+          slides: [{ column_blocks: [[textBlock('<p>Visible slide copy.</p>')]] }],
+        },
+      },
+    ]);
+
+    expect(document.text).toBe('Visible slide copy.');
+  });
+
+  it('still reads a legacy hero block, which stores its children in slides', () => {
+    const document = buildPageSeoDocument([
+      {
+        block_type: 'hero',
+        content: { slides: [{ column_blocks: [[headingBlock(1, 'Legacy Hero')]] }] },
+      },
+    ]);
+
+    expect(document.headings).toEqual([{ level: 1, order: 0, text: 'Legacy Hero' }]);
+  });
+});
+
+describe('buildPageSeoDocument on rich-text blocks', () => {
+  it('reads an html_content that holds a JSON-stringified Tiptap document', () => {
+    // Both storage formats are live in the same column; the app tells them apart
+    // by a leading brace, and so must anything that reads the column.
+    const tiptap = JSON.stringify({
+      content: [
+        { attrs: { level: 2 }, content: [{ text: 'Tiptap Heading', type: 'text' }], type: 'heading' },
+        { content: [{ text: 'Tiptap paragraph copy.', type: 'text' }], type: 'paragraph' },
+      ],
+      type: 'doc',
+    });
+
+    const document = buildPageSeoDocument([textBlock(tiptap)]);
+
+    expect(document.headings).toEqual([{ level: 2, order: 0, text: 'Tiptap Heading' }]);
+    expect(document.text).toBe('Tiptap Heading Tiptap paragraph copy.');
+  });
+
+  it('collects images and links embedded in prose', () => {
+    const document = buildPageSeoDocument([
+      textBlock('<p><img src="/inline.png" alt=""><a href="https://example.com">Away</a></p>'),
+    ]);
+
+    expect(document.images).toEqual([{ alt: '', src: '/inline.png' }]);
+    expect(document.links).toEqual([
+      { external: true, href: 'https://example.com', text: 'Away' },
+    ]);
+  });
+});
+
+describe('buildPageSeoDocument on image, button and video blocks', () => {
+  it('reports an image block through the same alt-text channel as inline images', () => {
+    const document = buildPageSeoDocument([
+      { block_type: 'image', content: { alt_text: 'A bowl of cat food', object_key: 'cat.png' } },
+      { block_type: 'image', content: { alt_text: '', external_url: 'https://cdn.test/b.png' } },
+    ]);
+
+    expect(document.images).toEqual([
+      { alt: 'A bowl of cat food', src: 'cat.png' },
+      { alt: '', src: 'https://cdn.test/b.png' },
+    ]);
+  });
+
+  it('prefers the external URL, which is what the renderer draws', () => {
+    const document = buildPageSeoDocument([
+      {
+        block_type: 'image',
+        content: { alt_text: 'Stock', external_url: 'https://cdn.test/a.png', object_key: 'a.png' },
+      },
+    ]);
+
+    expect(document.images[0]?.src).toBe('https://cdn.test/a.png');
+  });
+
+  it('ignores an image block with no source at all', () => {
+    // That block renders a "media not selected" placeholder, so flagging its
+    // missing alt text would be a complaint about an image nobody can see.
+    const document = buildPageSeoDocument([{ block_type: 'image', content: { alt_text: '' } }]);
+
+    expect(document.images).toEqual([]);
+  });
+
+  it('includes an image caption in the page text', () => {
+    const document = buildPageSeoDocument([
+      {
+        block_type: 'image',
+        content: { alt_text: 'Beans', caption: 'Beans drying in the sun.', object_key: 'b.png' },
+      },
+    ]);
+
+    expect(document.text).toBe('Beans drying in the sun.');
+  });
+
+  it('records a button as a link and its label as text', () => {
+    const document = buildPageSeoDocument([
+      { block_type: 'button', content: { text: 'Shop now', url: '/shop' } },
+    ]);
+
+    expect(document.links).toEqual([{ external: false, href: '/shop', text: 'Shop now' }]);
+    expect(document.text).toBe('Shop now');
+  });
+
+  it('keeps a button label even when no URL has been set yet', () => {
+    const document = buildPageSeoDocument([{ block_type: 'button', content: { text: 'Shop now' } }]);
+
+    expect(document.links).toEqual([]);
+    expect(document.text).toBe('Shop now');
+  });
+
+  it('takes a video block title as text when there is one', () => {
+    const document = buildPageSeoDocument([
+      { block_type: 'video_embed', content: { title: 'How we roast', url: 'https://v.test/1' } },
+      { block_type: 'video_embed', content: { url: 'https://v.test/2' } },
+    ]);
+
+    expect(document.text).toBe('How we roast');
+  });
+
+  it('takes a testimonial quote and its attribution', () => {
+    const document = buildPageSeoDocument([
+      {
+        block_type: 'testimonial',
+        content: { author_name: 'Jane Doe', author_title: 'CEO', quote: 'It changed my morning.' },
+      },
+    ]);
+
+    expect(document.text).toBe('It changed my morning. Jane Doe CEO');
+  });
+});
+
+describe('buildPageSeoDocument on content it cannot read', () => {
+  it('skips unknown and custom block types instead of guessing at their fields', () => {
+    const document = buildPageSeoDocument([
+      { block_type: 'pricing-table-v2', content: { internal_label: 'Do not index me', rows: 4 } },
+      { block_type: 'posts_grid', content: { columns: 3, postsPerPage: 6 } },
+      headingBlock(2, 'Real Heading'),
+    ]);
+
+    expect(document.headings).toEqual([{ level: 2, order: 0, text: 'Real Heading' }]);
+    expect(document.text).toBe('Real Heading');
+  });
+
+  it('survives every field being the wrong type or absent', () => {
+    const malformed: unknown[] = [
+      null,
+      undefined,
+      42,
+      'a string where a row should be',
+      [],
+      { block_type: 'text' },
+      { block_type: 'text', content: null },
+      { block_type: 'text', content: { html_content: 12 } },
+      { block_type: 'heading', content: {} },
+      { block_type: 'image', content: { alt_text: null, object_key: null } },
+      { block_type: 'button', content: { text: null, url: null } },
+      { block_type: 'section', content: { column_blocks: 'not an array' } },
+      { block_type: 'section', content: { column_blocks: [null, [null, 7]] } },
+      { block_type: 'section', content: {} },
+      { content: { html_content: '<p>No block type.</p>' } },
+    ];
+
+    expect(() => buildPageSeoDocument(malformed)).not.toThrow();
+
+    const document = buildPageSeoDocument(malformed);
+
+    // The one heading with no fields at all still surfaces, because an empty
+    // heading is a real defect the audit reports rather than a row to discard.
+    expect(document.headings).toEqual([{ level: 2, order: 0, text: '' }]);
+    expect(document.text).toBe('');
+    expect(document.words).toEqual([]);
+  });
+
+  it('stops descending once the nesting is deeper than any real page', () => {
+    // Guards against a corrupt row that refers back into itself: the walk is
+    // bounded, so the editor degrades to missing copy rather than to a blown
+    // stack while someone is typing.
+    let nested: unknown = textBlock('<p>Deepest copy.</p>');
+    for (let level = 0; level < 40; level += 1) {
+      nested = sectionBlock([[nested]]);
+    }
+
+    expect(() => buildPageSeoDocument([nested])).not.toThrow();
+    expect(buildPageSeoDocument([nested]).text).toBe('');
+  });
+});
+
+describe('buildPageSeoDocument feeding the page-level audit', () => {
+  it('no longer reports a missing H1 when the H1 lives in its own block', () => {
+    // This is the bug that prompted the page-level audit: the paragraph block
+    // was graded on its own and reported "this page has no H1" while the H1 sat
+    // one row above it, in a block the auditor was never shown.
+    const blocks = [
+      headingBlock(1, 'Fresh Cat Food'),
+      sectionBlock([[textBlock('<p>We ship every order fast and pack it with care.</p>')]]),
+    ];
+
+    const paragraphOnly = auditSeo({ document: buildPageSeoDocument([blocks[1]]) });
+    const wholePage = auditSeo({ document: buildPageSeoDocument(blocks) });
+
+    expect(paragraphOnly.issues.map((issue) => issue.id)).toContain('headings-missing-h1');
+    expect(wholePage.issues.map((issue) => issue.id)).not.toContain('headings-missing-h1');
+    expect(wholePage.headings).toEqual([{ level: 1, order: 0, text: 'Fresh Cat Food' }]);
+  });
+
+  it('counts the words of the whole page rather than of one paragraph', () => {
+    const sentence = '<p>We ship every order fast and we pack it with care for your pet today.</p>';
+    const blocks = new Array(30).fill(null).map(() => textBlock(sentence));
+
+    const oneBlock = auditSeo({ document: buildPageSeoDocument([blocks[0]]) });
+    const wholePage = auditSeo({ document: buildPageSeoDocument(blocks) });
+
+    expect(oneBlock.issues.map((issue) => issue.id)).toContain('content-thin');
+    expect(wholePage.issues.map((issue) => issue.id)).not.toContain('content-thin');
+  });
+});