fix(cloud): detect and ignore venv (#16056)

Co-authored-by: Ethan Harris <[email protected]>

(cherry picked from commit 3b323c8)

Author: yurijmikhalevich

src/lightning_app/runners/cloud.py

@@ -216,6 +216,15 @@ def dispatch(
         root = Path(self.entrypoint_file).absolute().parent
         cleanup_handle = _prepare_lightning_wheels_and_requirements(root)
         self.app._update_index_file()
+
+        # Create a default dotignore if it doesn't exist
+        if not (root / DOT_IGNORE_FILENAME).is_file():
+            with open(root / DOT_IGNORE_FILENAME, "w") as f:
+                f.write("venv/\n")
+                if (root / "bin" / "activate").is_file() or (root / "pyvenv.cfg").is_file():
+                    # the user is developing inside venv
+                    f.write("bin/\ninclude/\nlib/\npyvenv.cfg\n")
+
         repo = LocalSourceCodeDir(path=root)
         self._check_uploaded_folder(root, repo)
         requirements_file = root / "requirements.txt"

tests/tests_app/runners/test_cloud.py

@@ -1488,6 +1488,56 @@ def run(self):
         _validate_build_spec_and_compute(Work())
 
 
+def test_default_lightningignore(monkeypatch, caplog, tmpdir):
+    mock_client = mock.MagicMock()
+    mock_client.projects_service_list_memberships.return_value = V1ListMembershipsResponse(
+        memberships=[V1Membership(name="test-project", project_id="test-project-id")]
+    )
+    mock_client.lightningapp_instance_service_list_lightningapp_instances.return_value = (
+        V1ListLightningappInstancesResponse(lightningapps=[])
+    )
+    mock_client.lightningapp_v2_service_create_lightningapp_release.return_value = V1LightningappRelease(
+        cluster_id="test"
+    )
+    cloud_backend = mock.MagicMock(client=mock_client)
+    monkeypatch.setattr(backends, "CloudBackend", mock.MagicMock(return_value=cloud_backend))
+
+    class MyWork(LightningWork):
+        def run(self):
+            pass
+
+    app = LightningApp(MyWork())
+
+    path = Path(tmpdir)
+    cloud_runtime = cloud.CloudRuntime(app=app, entrypoint_file=path / "entrypoint.py")
+    monkeypatch.setattr(LocalSourceCodeDir, "upload", mock.MagicMock())
+
+    # write some files
+    write_file_of_size(path / "a.txt", 5 * 1000 * 1000)
+    write_file_of_size(path / "venv" / "foo.txt", 4 * 1000 * 1000)
+
+    assert not (path / ".lightningignore").exists()
+
+    with mock.patch(
+        "lightning_app.runners.cloud._parse_lightningignore", wraps=_parse_lightningignore
+    ) as parse_mock, mock.patch(
+        "lightning_app.source_code.local._copytree", wraps=_copytree
+    ) as copy_mock, caplog.at_level(
+        logging.WARN
+    ):
+        cloud_runtime.dispatch()
+
+    parse_mock.assert_called_once_with(())
+    assert copy_mock.mock_calls[0].kwargs["ignore_functions"][0].args[1] == set()
+
+    assert (path / ".lightningignore").exists()
+
+    assert f"Your application folder '{path.absolute()}' is more than 2 MB" in caplog.text
+    assert "The total size is 5.0 MB" in caplog.text
+    assert "2 files were uploaded"  # a.txt and .lightningignore
+    assert "files:\n5.0 MB: a.txt\nPerhaps" in caplog.text  # only this file appears
+
+
 @pytest.mark.parametrize(
     "lightning_app_instance, lightning_cloud_url, expected_url",
     [