obsidian_ollama/tests/auto-organizer.test.ts

import {
  AutoTagger,
  AutoLinker,
  normalizeTag,
  buildTagVocabulary,
  normalizeTagsAgainstVocabulary,
} from '../src/auto-organizer';
import { OllamaClient } from '../src/ollama-client';

// Mock dependencies
jest.mock('../src/ollama-client');
jest.mock('../src/utils', () => ({
  Logger: {
    info: jest.fn(),
    warn: jest.fn(),
    error: jest.fn(),
  },
}));

// Mock Obsidian
const mockModify = jest.fn();
const mockCachedRead = jest.fn();
const mockRead = jest.fn();
const mockGetMarkdownFiles = jest.fn();

const createMockVault = () => ({
  getMarkdownFiles: mockGetMarkdownFiles,
  cachedRead: mockCachedRead,
  read: mockRead,
  modify: mockModify,
});

const createMockApp = () => ({
  metadataCache: {
    getFileCache: jest.fn().mockReturnValue(null),
  },
});

describe('AutoTagger', () => {
  let tagger: AutoTagger;
  let mockVault: ReturnType<typeof createMockVault>;
  let mockApp: ReturnType<typeof createMockApp>;

  beforeEach(() => {
    jest.clearAllMocks();
    mockVault = createMockVault();
    mockApp = createMockApp();
    tagger = new AutoTagger(mockVault as any, mockApp as any, 'http://localhost:11434', 'llama3', {
      enabled: true,
      maxTagsPerNote: 5,
      minNoteLength: 50,
      maxNoteLength: 8000,
      tagPromptTemplate: 'Tags for {{title}}: {{content}}',
      dryRun: false,
      targetFolder: '',
      normalizeTags: true,
    });
  });

  describe('getUntaggedNotes', () => {
    it('should return files without frontmatter', async () => {
      const files = [{ path: 'note1.md' }, { path: 'note2.md' }] as any[];
      mockGetMarkdownFiles.mockReturnValue(files);
      mockApp.metadataCache.getFileCache
        .mockReturnValueOnce(null)
        .mockReturnValueOnce({ frontmatter: { tags: 'existing' } });

      const result = await tagger.getUntaggedNotes();
      expect(result).toHaveLength(1);
      expect(result[0].path).toBe('note1.md');
    });

    it('should return files with empty tags', async () => {
      const files = [{ path: 'note1.md' }] as any[];
      mockGetMarkdownFiles.mockReturnValue(files);
      mockApp.metadataCache.getFileCache.mockReturnValue({ frontmatter: { tags: '' } });

      const result = await tagger.getUntaggedNotes();
      expect(result).toHaveLength(1);
    });

    it('should return files with empty array tags', async () => {
      const files = [{ path: 'note1.md' }] as any[];
      mockGetMarkdownFiles.mockReturnValue(files);
      mockApp.metadataCache.getFileCache.mockReturnValue({ frontmatter: { tags: [] } });

      const result = await tagger.getUntaggedNotes();
      expect(result).toHaveLength(1);
    });

    it('should skip files with existing tags', async () => {
      const files = [{ path: 'note1.md' }] as any[];
      mockGetMarkdownFiles.mockReturnValue(files);
      mockApp.metadataCache.getFileCache.mockReturnValue({ frontmatter: { tags: ['ai', 'ml'] } });

      const result = await tagger.getUntaggedNotes();
      expect(result).toHaveLength(0);
    });

    it('should respect targetFolder', async () => {
      tagger.updateConfig({ ...(tagger as any).config, targetFolder: 'Projects' });
      const files = [
        { path: 'Projects/note1.md' },
        { path: 'Archive/note2.md' },
        { path: 'Projects/Sub/note3.md' },
      ] as any[];
      mockGetMarkdownFiles.mockReturnValue(files);
      mockApp.metadataCache.getFileCache.mockReturnValue(null);

      const result = await tagger.getUntaggedNotes();
      expect(result).toHaveLength(2);
      expect(result.map((f: any) => f.path)).toEqual([
        'Projects/note1.md',
        'Projects/Sub/note3.md',
      ]);
    });
  });

  describe('parseTagResponse', () => {
    it('should parse comma-separated tags', () => {
      const tags = (tagger as any).parseTagResponse('ai, machine-learning, obsidian');
      expect(tags).toEqual(['ai', 'machine-learning', 'obsidian']);
    });

    it('should clean up hashtags and quotes', () => {
      const tags = (tagger as any).parseTagResponse('#ai, "machine learning", #obsidian');
      expect(tags).toEqual(['ai', 'machine learning', 'obsidian']);
    });

    it('should limit to max tags', () => {
      const tags = (tagger as any).parseTagResponse('a, b, c, d, e, f, g');
      expect(tags).toHaveLength(5);
    });

    it('should filter empty tags', () => {
      const tags = (tagger as any).parseTagResponse('ai,, , ml');
      expect(tags).toEqual(['ai', 'ml']);
    });
  });

  describe('applyTags', () => {
    it('should add frontmatter to note without it', async () => {
      const file = { path: 'note.md' } as any;
      mockRead.mockResolvedValue('Just content');
      mockApp.metadataCache.getFileCache.mockReturnValue(null);

      await tagger.applyTags(file, ['ai', 'ml']);

      expect(mockModify).toHaveBeenCalledWith(file, '---\ntags: ai, ml\n---\n\nJust content');
    });

    it('should update existing frontmatter with tags', async () => {
      const file = { path: 'note.md' } as any;
      mockRead.mockResolvedValue('---\ndate: 2024-01-01\n---\nContent');
      mockApp.metadataCache.getFileCache.mockReturnValue({ frontmatter: {} });

      await tagger.applyTags(file, ['ai']);

      expect(mockModify).toHaveBeenCalledWith(file, expect.stringContaining('tags: ai'));
    });

    it('should replace existing tags line', async () => {
      const file = { path: 'note.md' } as any;
      mockRead.mockResolvedValue('---\ntags: old\n---\nContent');
      mockApp.metadataCache.getFileCache.mockReturnValue({ frontmatter: { tags: 'old' } });

      await tagger.applyTags(file, ['new']);

      expect(mockModify).toHaveBeenCalledWith(file, expect.stringContaining('tags: new'));
      expect(mockModify).not.toHaveBeenCalledWith(file, expect.stringContaining('tags: old'));
    });
  });

  describe('run', () => {
    it('should return early when disabled', async () => {
      tagger.updateConfig({ ...tagger['config'], enabled: false });
      const result = await tagger.run();
      expect(result.tagged).toBe(0);
    });

    it('should skip notes that are too short', async () => {
      const files = [{ path: 'note.md' }] as any[];
      mockGetMarkdownFiles.mockReturnValue(files);
      mockApp.metadataCache.getFileCache.mockReturnValue(null);
      mockRead.mockResolvedValue('Short');

      const result = await tagger.run();
      expect(result.skipped).toBeGreaterThanOrEqual(0);
    });

    it('should return dry-run proposals when dryRun is enabled', async () => {
      tagger.updateConfig({ ...tagger['config'], dryRun: true });
      const files = [{ path: 'note.md', basename: 'Note' }] as any[];
      mockGetMarkdownFiles.mockReturnValue(files);
      mockApp.metadataCache.getFileCache.mockReturnValue(null);
      mockRead.mockResolvedValue('A longer note about AI and machine learning with lots of interesting content that exceeds the minimum length requirement for tagging.');

      // Mock the ollamaClient on the tagger instance
      (tagger as any).ollamaClient = {
        chat: jest.fn().mockResolvedValue({
          content: 'ai, machine-learning',
          role: 'assistant',
        }),
      };

      const result = await tagger.run();
      expect(result.dryRun).toBeDefined();
      expect(result.dryRun!.length).toBe(1);
      expect(result.dryRun![0].proposedTags).toContain('ai');
      expect(result.dryRun![0].proposedTags).toContain('machine-learning');
    });
  });

  describe('normalizeTag', () => {
    it('should lowercase and hyphenate tags', () => {
      expect(normalizeTag('Machine Learning')).toBe('machine-learning');
      expect(normalizeTag('AI')).toBe('ai');
      expect(normalizeTag('obsidian-plugin')).toBe('obsidian-plugin');
    });

    it('should strip special characters', () => {
      expect(normalizeTag('C++')).toBe('c');
      expect(normalizeTag('Node.js')).toBe('nodejs');
    });

    it('should trim dashes', () => {
      expect(normalizeTag('-leading')).toBe('leading');
      expect(normalizeTag('trailing-')).toBe('trailing');
    });
  });

  describe('buildTagVocabulary', () => {
    it('should collect existing tags from vault frontmatter', () => {
      const mockFiles = [{ path: 'a.md' }, { path: 'b.md' }] as any[];
      const vault = createMockVault();
      vault.getMarkdownFiles.mockReturnValue(mockFiles);
      ;(vault as any).app = {
        metadataCache: {
          getFileCache: jest.fn().mockImplementation((f: any) => {
            if (f.path === 'a.md') return { frontmatter: { tags: ['machine-learning', 'ai'] } };
            if (f.path === 'b.md') return { frontmatter: { tags: 'obsidian-plugin' } };
            return null;
          }),
        },
      };

      const vocab = buildTagVocabulary(vault as any, (vault as any).app);
      expect(vocab.get('machine-learning')).toBe('machine-learning');
      expect(vocab.get('ai')).toBe('ai');
      expect(vocab.get('obsidian-plugin')).toBe('obsidian-plugin');
    });
  });

  describe('normalizeTagsAgainstVocabulary', () => {
    it('should prefer canonical forms from vocabulary', () => {
      const vocab = new Map([
        ['machine-learning', 'machine-learning'],
        ['obsidian', 'Obsidian'],
      ]);
      const result = normalizeTagsAgainstVocabulary(
        ['Machine Learning', 'obsidian', 'new-tag'],
        vocab
      );
      expect(result).toContain('machine-learning');
      expect(result).toContain('Obsidian');
      expect(result).toContain('new-tag');
    });

    it('should deduplicate normalized tags', () => {
      const vocab = new Map();
      const result = normalizeTagsAgainstVocabulary(['ai', 'AI', 'Ai'], vocab);
      expect(result).toEqual(['ai']);
    });
  });
});

describe('AutoLinker', () => {
  let linker: AutoLinker;
  let mockVault: ReturnType<typeof createMockVault>;
  let mockIndexer: { searchVault: jest.Mock };

  beforeEach(() => {
    jest.clearAllMocks();
    mockVault = createMockVault();
    mockIndexer = {
      searchVault: jest.fn(),
    };

    linker = new AutoLinker(mockVault as any, mockIndexer as any, {
      enabled: true,
      maxLinksPerNote: 3,
      similarityThreshold: 0.5,
    });
  });

  describe('findRelatedNotes', () => {
    it('should return related notes excluding self', async () => {
      const file = { path: 'note.md', basename: 'Note' } as any;
      mockVault.read.mockResolvedValue('Content about AI');
      mockIndexer.searchVault!.mockResolvedValue([
        { path: 'other.md', title: 'Other', score: 0.9, content: '' },
        { path: 'note.md', title: 'Note', score: 0.95, content: '' },
      ]);

      const result = await linker.findRelatedNotes(file);
      expect(result).toHaveLength(1);
      expect(result[0].path).toBe('other.md');
    });

    it('should filter by similarity threshold', async () => {
      const file = { path: 'note.md', basename: 'Note' } as any;
      mockVault.read.mockResolvedValue('Content');
      mockIndexer.searchVault!.mockResolvedValue([
        { path: 'high.md', title: 'High', score: 0.8, content: '' },
        { path: 'low.md', title: 'Low', score: 0.3, content: '' },
      ]);

      const result = await linker.findRelatedNotes(file);
      expect(result).toHaveLength(1);
      expect(result[0].path).toBe('high.md');
    });
  });

  describe('addRelatedLinks', () => {
    it('should add Related Notes section', async () => {
      const file = { path: 'note.md' } as any;
      mockVault.read.mockResolvedValue('# Note\n\nContent');

      await linker.addRelatedLinks(file, [{ path: 'other.md', title: 'Other' }]);

      expect(mockVault.modify).toHaveBeenCalledWith(
        file,
        expect.stringContaining('## Related Notes')
      );
      expect(mockVault.modify).toHaveBeenCalledWith(
        file,
        expect.stringContaining('[[Other|other]]')
      );
    });

    it('should skip if Related Notes already exists', async () => {
      const file = { path: 'note.md' } as any;
      mockVault.read.mockResolvedValue('# Note\n\n## Related Notes\nAlready linked');

      await linker.addRelatedLinks(file, [{ path: 'other.md', title: 'Other' }]);
      expect(mockVault.modify).not.toHaveBeenCalled();
    });
  });

  describe('run', () => {
    it('should return early when disabled', async () => {
      linker.updateConfig({ enabled: false, maxLinksPerNote: 3, similarityThreshold: 0.5 });
      const result = await linker.run();
      expect(result.linked).toBe(0);
    });

    it('should respect targetFolder', async () => {
      linker.setTargetFolder('Projects');
      const files = [{ path: 'Projects/note1.md' }, { path: 'Archive/note2.md' }] as any[];
      mockVault.getMarkdownFiles.mockReturnValue(files);
      mockVault.read.mockResolvedValue('Content');
      mockIndexer.searchVault!.mockResolvedValue([]);

      const result = await linker.run();
      expect(mockVault.read).toHaveBeenCalledTimes(1);
      expect(mockVault.read).toHaveBeenCalledWith(files[0]);
    });

    it('should update targetFolder through updateConfig', async () => {
      linker.setTargetFolder('Projects');
      linker.updateConfig({
        enabled: true,
        maxLinksPerNote: 3,
        similarityThreshold: 0.5,
        targetFolder: 'Archive',
      });
      const files = [{ path: 'Projects/note1.md' }, { path: 'Archive/note2.md' }] as any[];
      mockVault.getMarkdownFiles.mockReturnValue(files);
      mockVault.read.mockResolvedValue('Content');
      mockIndexer.searchVault!.mockResolvedValue([]);

      await linker.run();

      expect(mockVault.read).toHaveBeenCalledTimes(1);
      expect(mockVault.read).toHaveBeenCalledWith(files[1]);
    });

    it('should return dry-run proposals when dryRun is true', async () => {
      const files = [{ path: 'note.md' }] as any[];
      mockVault.getMarkdownFiles.mockReturnValue(files);
      mockVault.read.mockResolvedValue('Content');
      mockIndexer.searchVault!.mockResolvedValue([
        { path: 'other.md', title: 'Other', score: 0.9, content: '' },
      ]);

      const result = await linker.run(true);
      expect(result.dryRun).toBeDefined();
      expect(result.dryRun!.length).toBe(1);
      expect(result.dryRun![0].relatedNotes[0].path).toBe('other.md');
      expect(mockVault.modify).not.toHaveBeenCalled();
    });
  });
});