broccolini-bot/tests/gmailAttachments.test.js

/**
 * fetchMessageAttachments — Gmail attachment → discord.js file descriptor tests.
 *
 * Uses a fake gmail client (no module mocking) so we exercise the real
 * collectAttachmentParts walk, the base64url→Buffer decode, the size ceiling,
 * the 10-file cap, and the best-effort skip-on-failure behavior.
 */
import { describe, it, expect, vi } from 'vitest';

import { collectAttachmentParts, fetchMessageAttachments } from '../services/gmail.js';

// Build a payload part carrying a real attachment (filename + attachmentId).
function attPart(filename, attachmentId, size, mimeType = 'image/png') {
  return { filename, mimeType, body: { attachmentId, size } };
}

// Fake gmail whose attachments.get returns base64url of the given string per id.
function fakeGmail(dataById) {
  return {
    users: {
      messages: {
        attachments: {
          get: vi.fn(async ({ id }) => {
            if (!(id in dataById)) throw new Error('not found');
            const b64url = Buffer.from(dataById[id]).toString('base64')
              .replace(/\+/g, '-').replace(/\//g, '_').replace(/=+$/, '');
            return { data: { data: b64url } };
          })
        }
      }
    }
  };
}

describe('collectAttachmentParts', () => {
  it('finds attachment parts at any nesting depth, skips inline text', () => {
    const payload = {
      parts: [
        { mimeType: 'text/plain', body: { data: 'aGk=' } },
        { mimeType: 'multipart/mixed', parts: [attPart('log.txt', 'A1', 12)] }
      ]
    };
    const parts = collectAttachmentParts(payload);
    expect(parts).toHaveLength(1);
    expect(parts[0]).toMatchObject({ filename: 'log.txt', attachmentId: 'A1', size: 12 });
  });

  it('carries over an embedded screenshot that has no filename', () => {
    const payload = {
      parts: [
        { mimeType: 'image/png', body: { attachmentId: 'CID1', size: 40000 } }
      ]
    };
    const parts = collectAttachmentParts(payload);
    expect(parts).toHaveLength(1);
    expect(parts[0]).toMatchObject({ filename: 'screenshot-1.png', attachmentId: 'CID1' });
  });

  it('skips a nameless text/html body served as an attachmentId part', () => {
    const payload = {
      parts: [
        { mimeType: 'text/html', body: { attachmentId: 'BODY', size: 200000 } }
      ]
    };
    expect(collectAttachmentParts(payload)).toEqual([]);
  });

  it('names a nameless non-image attachment "attachment-N"', () => {
    const payload = {
      parts: [
        { mimeType: 'application/pdf', body: { attachmentId: 'P1', size: 1000 } }
      ]
    };
    expect(collectAttachmentParts(payload)[0].filename).toBe('attachment-1.pdf');
  });
});

describe('fetchMessageAttachments', () => {
  it('decodes base64url attachment data into Buffers', async () => {
    const payload = { parts: [attPart('hello.txt', 'A1', 5)] };
    const gmail = fakeGmail({ A1: 'hello' });

    const { files, skipped } = await fetchMessageAttachments('msg1', payload, gmail);

    expect(skipped).toEqual([]);
    expect(files).toHaveLength(1);
    expect(files[0].name).toBe('hello.txt');
    expect(Buffer.isBuffer(files[0].attachment)).toBe(true);
    expect(files[0].attachment.toString()).toBe('hello');
  });

  it('skips parts over the 25 MB ceiling without fetching them', async () => {
    const payload = { parts: [attPart('huge.bin', 'BIG', 26 * 1024 * 1024)] };
    const gmail = fakeGmail({ BIG: 'x' });

    const { files, skipped } = await fetchMessageAttachments('msg1', payload, gmail);

    expect(files).toEqual([]);
    expect(skipped).toEqual(['huge.bin']);
    expect(gmail.users.messages.attachments.get).not.toHaveBeenCalled();
  });

  it('records a failed download as skipped, keeps the rest', async () => {
    const payload = { parts: [attPart('ok.txt', 'A1', 2), attPart('gone.txt', 'MISSING', 2)] };
    const gmail = fakeGmail({ A1: 'ok' });

    const { files, skipped } = await fetchMessageAttachments('msg1', payload, gmail);

    expect(files.map(f => f.name)).toEqual(['ok.txt']);
    expect(skipped).toEqual(['gone.txt']);
  });

  it('caps at 10 files, skipping the overflow', async () => {
    const data = {};
    const parts = [];
    for (let i = 0; i < 12; i++) {
      const id = `A${i}`;
      data[id] = `f${i}`;
      parts.push(attPart(`file${i}.txt`, id, 2));
    }
    const { files, skipped } = await fetchMessageAttachments('msg1', { parts }, fakeGmail(data));

    expect(files).toHaveLength(10);
    expect(skipped).toHaveLength(2);
  });

  it('sanitizes CRLF and backticks out of filenames', async () => {
    const payload = { parts: [attPart('bad\nname`.txt', 'A1', 2)] };
    const gmail = fakeGmail({ A1: 'hi' });

    const { files } = await fetchMessageAttachments('msg1', payload, gmail);

    expect(files[0].name).toBe('bad name .txt');
  });
});