obsidian_ollama/tests/vault-indexer.test.ts

import { VaultIndexer } from '../src/vault-indexer';
import { VaultIndexEntry } from '../src/types';

// Mock Obsidian types
interface MockTFile {
  basename: string;
  path: string;
}

interface MockVault {
  getMarkdownFiles: () => MockTFile[];
  read: (file: MockTFile) => Promise<string>;
}

describe('VaultIndexer', () => {
  let indexer: VaultIndexer;
  let mockVault: MockVault;

  beforeEach(() => {
    mockVault = {
      getMarkdownFiles: jest.fn().mockReturnValue([]),
      read: jest.fn(),
    };
    indexer = new VaultIndexer(mockVault as unknown as any);
    jest.clearAllMocks();
  });

  describe('searchVault', () => {
    it('should return empty array when no files exist', async () => {
      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([]);
      const results = await indexer.searchVault('test', 5);
      expect(results).toEqual([]);
    });

    it('should return empty array for empty or whitespace-only query', async () => {
      const file: MockTFile = { basename: 'test', path: 'test.md' };
      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file]);
      mockVault.read = jest.fn().mockResolvedValue('some content');

      const results1 = await indexer.searchVault('', 5);
      const results2 = await indexer.searchVault('   ', 5);
      expect(results1).toEqual([]);
      expect(results2).toEqual([]);
    });

    it('should return files matching the query', async () => {
      const file1: MockTFile = { basename: 'notes', path: 'notes.md' };
      const file2: MockTFile = { basename: 'todo', path: 'todo.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'notes') {
          return 'These are my important notes about programming algorithms';
        }
        return 'Buy milk and eggs';
      });

      const results = await indexer.searchVault('programming', 5);
      expect(results.length).toBe(1);
      expect(results[0].title).toBe('notes');
      expect(results[0].score).toBeGreaterThan(0);
    });

    it('should respect the limit parameter', async () => {
      const files: MockTFile[] = [];
      for (let i = 0; i < 10; i++) {
        files.push({ basename: `file${i}`, path: `file${i}.md` });
      }

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue(files);
      mockVault.read = jest.fn().mockResolvedValue('important keyword test');

      const results = await indexer.searchVault('keyword', 3);
      expect(results.length).toBeLessThanOrEqual(3);
    });

    it('should return results sorted by score descending', async () => {
      const file1: MockTFile = { basename: 'one', path: 'one.md' };
      const file2: MockTFile = { basename: 'two', path: 'two.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'one') {
          return 'keyword keyword keyword important';
        }
        return 'keyword';
      });

      const results = await indexer.searchVault('keyword', 5);
      if (results.length >= 2) {
        expect(results[0].score).toBeGreaterThanOrEqual(results[1].score);
      }
    });

    it('should truncate content previews to 500 characters', async () => {
      const file: MockTFile = { basename: 'long', path: 'long.md' };
      const longContent = 'content '.repeat(100); // Use meaningful words, not just 'a'

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file]);
      mockVault.read = jest.fn().mockResolvedValue(longContent);

      const results = await indexer.searchVault('content', 5);
      expect(results.length).toBeGreaterThan(0);
      expect(results[0].content.length).toBeLessThanOrEqual(500);
    });

    it('should process files in batches to handle large vaults', async () => {
      const files: MockTFile[] = [];
      for (let i = 0; i < 25; i++) {
        files.push({ basename: `file${i}`, path: `file${i}.md` });
      }

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue(files);
      mockVault.read = jest.fn().mockResolvedValue('important test keyword');

      const results = await indexer.searchVault('test', 5);
      expect(mockVault.read).toHaveBeenCalledTimes(25);
      expect(results.length).toBeGreaterThan(0);
    });

    it('should filter out files with zero score', async () => {
      const file1: MockTFile = { basename: 'match', path: 'match.md' };
      const file2: MockTFile = { basename: 'nomatch', path: 'nomatch.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'match') {
          return 'relevant keyword algorithm';
        }
        return 'nothing relevant here at all';
      });

      const results = await indexer.searchVault('keyword', 5);
      expect(results.length).toBe(1);
      expect(results[0].title).toBe('match');
    });

    it('should handle vault.read errors gracefully', async () => {
      const file1: MockTFile = { basename: 'good', path: 'good.md' };
      const file2: MockTFile = { basename: 'bad', path: 'bad.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'good') {
          return 'important keyword test';
        }
        throw new Error('Permission denied');
      });

      const consoleWarnSpy = jest.spyOn(console, 'warn').mockImplementation();

      const results = await indexer.searchVault('keyword', 5);
      expect(results.length).toBe(1);
      expect(results[0].title).toBe('good');
      expect(consoleWarnSpy).toHaveBeenCalledWith(expect.stringContaining('Permission denied'));

      consoleWarnSpy.mockRestore();
    });

    it('should give higher scores to title matches', async () => {
      const file1: MockTFile = { basename: 'algorithm', path: 'algorithm.md' };
      const file2: MockTFile = { basename: 'other', path: 'other.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'algorithm') {
          return 'Some other content here';
        }
        return 'This file discusses algorithm design patterns';
      });

      const results = await indexer.searchVault('algorithm', 5);
      expect(results.length).toBe(2);
      // File with title match should be first
      expect(results[0].title).toBe('algorithm');
    });

    it('should give higher scores to heading matches', async () => {
      const file1: MockTFile = { basename: 'file1', path: 'file1.md' };
      const file2: MockTFile = { basename: 'file2', path: 'file2.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'file1') {
          return '# Algorithm Design\n\nThis discusses design patterns';
        }
        return 'This file mentions algorithm somewhere in the body text';
      });

      const results = await indexer.searchVault('algorithm', 5);
      expect(results.length).toBe(2);
      // File with heading match should score higher
      expect(results[0].title).toBe('file1');
    });

    it('should give higher scores to frontmatter matches', async () => {
      const file1: MockTFile = { basename: 'file1', path: 'file1.md' };
      const file2: MockTFile = { basename: 'file2', path: 'file2.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'file1') {
          return '---\ntags: algorithm design\n---\n\nSome content here';
        }
        return 'This file mentions algorithm in the body';
      });

      const results = await indexer.searchVault('algorithm', 5);
      expect(results.length).toBe(2);
      // File with frontmatter match should score higher
      expect(results[0].title).toBe('file1');
    });

    it('should handle phrase matching with bonus', async () => {
      const file1: MockTFile = { basename: 'file1', path: 'file1.md' };
      const file2: MockTFile = { basename: 'file2', path: 'file2.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file1, file2]);
      mockVault.read = jest.fn().mockImplementation(async (file: MockTFile) => {
        if (file.basename === 'file1') {
          return 'This discusses the design pattern algorithm';
        }
        return 'This discusses design and pattern and algorithm separately';
      });

      const results = await indexer.searchVault('design pattern', 5);
      expect(results.length).toBe(2);
    });

    it('should filter out stop words from query', async () => {
      const file: MockTFile = { basename: 'test', path: 'test.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file]);
      mockVault.read = jest.fn().mockResolvedValue('important keyword here');

      // Query with stop words should still find the keyword
      const results = await indexer.searchVault('the important keyword', 5);
      expect(results.length).toBe(1);
      expect(results[0].title).toBe('test');
    });

    it('should handle files with no matching content', async () => {
      const file: MockTFile = { basename: 'test', path: 'test.md' };

      mockVault.getMarkdownFiles = jest.fn().mockReturnValue([file]);
      mockVault.read = jest.fn().mockResolvedValue('nothing relevant');

      const results = await indexer.searchVault('nonexistent', 5);
      expect(results.length).toBe(0);
    });
  });

  describe('tokenize', () => {
    it('should filter out stop words', () => {
      const tokens = (indexer as any).tokenize('the quick brown fox');
      expect(tokens).not.toContain('the');
      expect(tokens).toContain('quick');
      expect(tokens).toContain('brown');
      expect(tokens).toContain('fox');
    });

    it('should convert to lowercase', () => {
      const tokens = (indexer as any).tokenize('Hello WORLD');
      expect(tokens).toEqual(['hello', 'world']);
    });

    it('should handle punctuation', () => {
      const tokens = (indexer as any).tokenize('Hello, world!');
      expect(tokens).toEqual(['hello', 'world']);
    });

    it('should filter very short tokens', () => {
      const tokens = (indexer as any).tokenize('a b test word');
      expect(tokens).not.toContain('a');
      expect(tokens).not.toContain('b');
      expect(tokens).toContain('test');
      expect(tokens).toContain('word');
    });
  });

  describe('calculateWeightedScore', () => {
    it('should return 0 when no tokens match', () => {
      const content = 'important algorithm design';
      const queryTokens = (indexer as any).tokenize('nonexistent');
      const tokenized = (indexer as any).tokenizeContent(content, {
        basename: 'test',
        path: 'test.md',
      } as any);
      const score = (indexer as any).calculateWeightedScore(tokenized, queryTokens);
      expect(score.score).toBe(0);
    });

    it('should score higher when more tokens match', () => {
      const content = 'algorithm design pattern implementation';
      const tokenized = (indexer as any).tokenizeContent(content, {
        basename: 'test',
        path: 'test.md',
      } as any);
      const query1 = 'algorithm';
      const query2 = 'algorithm design pattern';
      const score1 = (indexer as any).calculateWeightedScore(
        tokenized,
        (indexer as any).tokenize(query1)
      );
      const score2 = (indexer as any).calculateWeightedScore(
        tokenized,
        (indexer as any).tokenize(query2)
      );
      expect(score2.score).toBeGreaterThan(score1.score);
    });

    it('should be case insensitive', () => {
      const content = 'Important Algorithm Design';
      const tokenized = (indexer as any).tokenizeContent(content, {
        basename: 'test',
        path: 'test.md',
      } as any);
      const query = 'important algorithm';
      const score = (indexer as any).calculateWeightedScore(
        tokenized,
        (indexer as any).tokenize(query)
      );
      expect(score.score).toBeGreaterThan(0);
    });

    it('should handle word boundary matching', () => {
      const content = 'algorithm';
      const tokenized = (indexer as any).tokenizeContent(content, {
        basename: 'test',
        path: 'test.md',
      } as any);
      const query = 'algorithm';
      const score = (indexer as any).calculateWeightedScore(
        tokenized,
        (indexer as any).tokenize(query)
      );
      expect(score.score).toBeGreaterThan(0);
    });
  });

  describe('exactMatch with stemming', () => {
    it('should match exact words', () => {
      expect((indexer as any).exactMatch('test', 'test')).toBe(true);
    });

    it('should match plurals', () => {
      expect((indexer as any).exactMatch('tests', 'test')).toBe(true);
      expect((indexer as any).exactMatch('test', 'tests')).toBe(true);
    });

    it('should handle -ed suffix', () => {
      expect((indexer as any).exactMatch('tested', 'test')).toBe(true);
    });

    it('should handle -ing suffix', () => {
      expect((indexer as any).exactMatch('testing', 'test')).toBe(true);
    });

    it('should not match unrelated words', () => {
      expect((indexer as any).exactMatch('apple', 'banana')).toBe(false);
    });
  });

  describe('tokenizeContent', () => {
    it('should extract headings from markdown', () => {
      const content = '# Heading 1\n\n# Heading 2\n\nSome content';
      const tokenized = (indexer as any).tokenizeContent(content, {
        basename: 'test',
        path: 'test.md',
      } as any);
      expect(tokenized.headings).toContain('Heading 1');
      expect(tokenized.headings).toContain('Heading 2');
    });

    it('should extract frontmatter', () => {
      const content = '---\ntags: algorithm\ntitle: test\n---\n\nSome content';
      const tokenized = (indexer as any).tokenizeContent(content, {
        basename: 'test',
        path: 'test.md',
      } as any);
      expect(tokenized.frontmatter.tags).toBe('algorithm');
      expect(tokenized.frontmatter.title).toBe('test');
    });

    it('should extract first paragraph', () => {
      const content = 'First paragraph here.\n\nSecond paragraph here.';
      const tokenized = (indexer as any).tokenizeContent(content, {
        basename: 'test',
        path: 'test.md',
      } as any);
      expect(tokenized.firstParagraph).toContain('First');
      expect(tokenized.firstParagraph).not.toContain('Second');
    });
  });
});