llmxml/src/parser/__tests__/llm.test.ts

Convert between markdown and LLM-friendly pseudo-XML

import type { TagNode, TextNode } from '../../types/ast';
import { describe, it, expect } from 'vitest';
import { parse } from '../llm';

describe('LLM Parser', () => {
  it('should parse code blocks', () => {
    const input = `<Section>
\`\`\`js
const x = 1;
\`\`\`
</Section>`;

    const { ast } = parse(input);
    expect(ast).toHaveLength(1);

    const section = ast[0] as TagNode;
    const codeBlock = section.children?.find(n => (n as TextNode).textType === 'code');
    expect(codeBlock).toBeDefined();
    expect((codeBlock as TextNode).language).toBe('js');
  });

  it('should preserve whitespace around code blocks', () => {
    const input = `<Section>
Text before
\`\`\`js
const x = 1;
\`\`\`
Text after
</Section>`;

    const { ast } = parse(input);
    expect(ast).toHaveLength(1);

    const section = ast[0] as TagNode;
    const children = section.children || [];
    expect(children.some(n => (n as TextNode).textType === 'code')).toBe(true);
    expect(children.some(n => !(n as TextNode).textType && (n as TextNode).value.includes('Text before'))).toBe(true);
    expect(children.some(n => !(n as TextNode).textType && (n as TextNode).value.includes('Text after'))).toBe(true);
  });

  it('requires capital letter for tag names', () => {
    const input = '<section></section>';
    const { errors } = parse(input);
    expect(errors).toHaveLength(1);
    expect(errors[0].message).toBe('Tag names must start with a capital letter');
  });

  it('should parse tag attributes', () => {
    const input = '<Section title="Test" hlevel="2">Content</Section>';
    const { ast } = parse(input);

    const section = ast[0] as TagNode;
    expect(section.attributes).toMatchObject({
      title: 'Test',
      hlevel: '2'
    });
  });

  it('should use tag name as title if not specified', () => {
    const input = '<Section hlevel="2">Content</Section>';
    const { ast } = parse(input);

    const section = ast[0] as TagNode;
    expect(section.attributes).toMatchObject({
      title: 'Section',
      hlevel: '2'
    });
  });

  describe('error handling', () => {
    it('detects mismatched tags', () => {
      const input = '<Section>Content</Other>';
      const { errors } = parse(input);
      expect(errors).toHaveLength(1);
      expect(errors[0].message).toContain('Mismatched tags');
    });

    it('handles unclosed tags', () => {
      const input = '<Section>Content';
      const { errors } = parse(input);
      expect(errors).toHaveLength(1);
      expect(errors[0].message).toContain('Mismatched tags');
    });

    it('provides location information', () => {
      const input = `<Section>
  <Other>
</Section>`;
      const { errors } = parse(input);
      expect(errors).toHaveLength(1);
      expect(errors[0].location).toBeDefined();
      expect(errors[0].location.start.line).toBe(2);
    });

    it('handles malformed code blocks', () => {
      const input = `<Section>
\`\`\`js
const x = 1;
</Section>`;
      const { errors } = parse(input);
      expect(errors).toHaveLength(1);
      expect(errors[0].message).toBe('Unclosed code block');
    });
  });
});