hotfix import torch (#15849)

* fix import torch

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* plugin

* fix

* skip

* patch require

* seed

* warn

* .

* ..

* skip True

* 0.0.3

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

.github/workflows/ci-pkg-install.yml

@@ -46,7 +46,7 @@ jobs:
       - name: DocTests actions
         working-directory: .actions/
         run: |
-          pip install pytest -q
+          pip install -q pytest
           python -m pytest setup_tools.py
 
       - run: python -c "print('NB_DIRS=' + str(2 if '${{ matrix.pkg-name }}' == 'pytorch' else 1))" >> $GITHUB_ENV
@@ -67,7 +67,10 @@ jobs:
 
       - name: DocTest package
         env:
+          LIGHTING_TESTING: 1  # path for require wrapper
           PY_IGNORE_IMPORTMISMATCH: 1
         run: |
+          pip install -q "pytest-doctestplus>=0.9.0"
+          pip list
           PKG_NAME=$(python -c "print({'app': 'lightning_app', 'lite': 'lightning_lite', 'pytorch': 'pytorch_lightning', 'lightning': 'lightning'}['${{matrix.pkg-name}}'])")
           python -m pytest src/${PKG_NAME} --ignore-glob="**/cli/*-template/**"

requirements/app/base.txt

@@ -12,4 +12,4 @@ beautifulsoup4>=4.8.0, <4.11.2
 inquirer>=2.10.0
 psutil<5.9.4
 click<=8.1.3
-lightning_api_access>=0.0.1
+lightning_api_access>=0.0.3

requirements/app/test.txt

@@ -3,6 +3,7 @@ codecov==2.1.12
 pytest==7.2.0
 pytest-timeout==2.1.0
 pytest-cov==4.0.0
+pytest-doctestplus>=0.9.0
 playwright==1.27.1
 httpx
 trio<0.22.0

src/lightning_app/components/serve/python_server.py

@@ -4,7 +4,6 @@
 from pathlib import Path
 from typing import Any, Dict, Optional
 
-import torch
 import uvicorn
 from fastapi import FastAPI
 from pydantic import BaseModel
@@ -13,16 +12,21 @@
 from lightning_app.core.queues import MultiProcessQueue
 from lightning_app.core.work import LightningWork
 from lightning_app.utilities.app_helpers import Logger
+from lightning_app.utilities.imports import _is_torch_available, requires
 from lightning_app.utilities.proxies import _proxy_setattr, unwrap, WorkRunExecutor, WorkStateObserver
 
 logger = Logger(__name__)
 
+# Skip doctests if requirements aren't available
+if not _is_torch_available():
+    __doctest_skip__ = ["PythonServer", "PythonServer.*"]
+
 
 class _PyTorchSpawnRunExecutor(WorkRunExecutor):
 
     """This Executor enables to move PyTorch tensors on GPU.
 
-    Without this executor, it woud raise the following expection:
+    Without this executor, it would raise the following exception:
     RuntimeError: Cannot re-initialize CUDA in forked subprocess.
     To use CUDA with multiprocessing, you must use the 'spawn' start method
     """
@@ -86,6 +90,7 @@ def _get_sample_data() -> Dict[Any, Any]:
 
 
 class PythonServer(LightningWork, abc.ABC):
+    @requires("torch")
     def __init__(  # type: ignore
         self,
         host: str = "127.0.0.1",
@@ -127,15 +132,16 @@ def predict(self, request):
                 and this can be accessed as `response.json()["prediction"]` in the client if
                 you are using requests library
 
-        .. doctest::
+        Example:
 
             >>> from lightning_app.components.serve.python_server import PythonServer
             >>> from lightning_app import LightningApp
-            >>>
             ...
             >>> class SimpleServer(PythonServer):
+            ...
             ...     def setup(self):
             ...         self._model = lambda x: x + " " + x
+            ...
             ...     def predict(self, request):
             ...         return {"prediction": self._model(request.image)}
             ...
@@ -199,11 +205,13 @@ def _get_sample_dict_from_datatype(datatype: Any) -> dict:
         return out
 
     def _attach_predict_fn(self, fastapi_app: FastAPI) -> None:
+        from torch import inference_mode
+
         input_type: type = self.configure_input_type()
         output_type: type = self.configure_output_type()
 
         def predict_fn(request: input_type):  # type: ignore
-            with torch.inference_mode():
+            with inference_mode():
                 return self.predict(request)
 
         fastapi_app.post("/predict", response_model=output_type)(predict_fn)

src/lightning_app/utilities/imports.py

@@ -12,8 +12,10 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """General utilities."""
+
 import functools
 import os
+import warnings
 from typing import List, Union
 
 from lightning_utilities.core.imports import module_available
@@ -52,10 +54,13 @@ def decorator(func):
         @functools.wraps(func)
         def wrapper(*args, **kwargs):
             unavailable_modules = [f"'{module}'" for module in module_paths if not module_available(module)]
-            if any(unavailable_modules) and not bool(int(os.getenv("LIGHTING_TESTING", "0"))):
-                raise ModuleNotFoundError(
-                    f"Required dependencies not available. Please run: pip install {' '.join(unavailable_modules)}"
-                )
+            if any(unavailable_modules):
+                is_lit_testing = bool(int(os.getenv("LIGHTING_TESTING", "0")))
+                msg = f"Required dependencies not available. Please run: pip install {' '.join(unavailable_modules)}"
+                if is_lit_testing:
+                    warnings.warn(msg)
+                else:
+                    raise ModuleNotFoundError(msg)
             return func(*args, **kwargs)
 
         return wrapper

src/lightning_app/utilities/name_generator.py

@@ -1332,12 +1332,13 @@ def get_unique_name():
     Original source:
         https://raw.githubusercontent.com/moby/moby/master/pkg/namesgenerator/names-generator.go
 
-    Examples
-    --------
-    >>> get_unique_name()  # doctest: +SKIP
-    'focused-turing-23'
-    >>> get_unique_name()  # doctest: +SKIP
-    'thirsty-allen-9200'
+    Examples:
+
+    >>> import random ; random.seed(42)
+    >>> get_unique_name()
+    'meek-ardinghelli-4506'
+    >>> get_unique_name()
+    'truthful-dijkstra-2286'
     """
     adjective, surname, i = choice(_adjectives), choice(_surnames), randint(0, 9999)
     return f"{adjective}-{surname}-{i}"

tests/tests_app/core/test_lightning_api.py

@@ -106,7 +106,7 @@ def run(self):
 
 
 # TODO: Find why this test is flaky.
-@pytest.mark.skipif(True, reason="flaky test.")
+@pytest.mark.skip(reason="flaky test.")
 @pytest.mark.parametrize("runtime_cls", [SingleProcessRuntime])
 def test_app_state_api_with_flows(runtime_cls, tmpdir):
     """This test validates the AppState can properly broadcast changes from flows."""
@@ -180,7 +180,7 @@ def maybe_apply_changes(self):
 
 
 # FIXME: This test doesn't assert anything
-@pytest.mark.skipif(True, reason="TODO: Resolve flaky test.")
+@pytest.mark.skip(reason="TODO: Resolve flaky test.")
 @pytest.mark.parametrize("runtime_cls", [SingleProcessRuntime, MultiProcessRuntime])
 def test_app_stage_from_frontend(runtime_cls):
     """This test validates that delta from the `api_delta_queue` manipulating the ['app_state']['stage'] would

tests/tests_app/core/test_lightning_app.py

@@ -582,7 +582,7 @@ def run(self):
 
 
 # TODO (tchaton) Resolve this test.
-@pytest.mark.skipif(True, reason="flaky test which never terminates")
+@pytest.mark.skip(reason="flaky test which never terminates")
 @pytest.mark.parametrize("runtime_cls", [MultiProcessRuntime])
 @pytest.mark.parametrize("use_same_args", [False, True])
 def test_state_wait_for_all_all_works(tmpdir, runtime_cls, use_same_args):

tests/tests_app/structures/test_structures.py

@@ -308,7 +308,7 @@ def run(self):
         self.counter += 1
 
 
-@pytest.mark.skipif(True, reason="tchaton: Resolve this test.")
+@pytest.mark.skip(reason="tchaton: Resolve this test.")
 @pytest.mark.parametrize("runtime_cls", [MultiProcessRuntime, SingleProcessRuntime])
 @pytest.mark.parametrize("run_once_iterable", [False, True])
 @pytest.mark.parametrize("cache_calls", [False, True])

tests/tests_app/utilities/packaging/test_docker.py

@@ -12,7 +12,7 @@
 from lightning_app.utilities.redis import check_if_redis_running
 
 
-@pytest.mark.skipif(True, reason="FIXME (tchaton)")
+@pytest.mark.skip(reason="FIXME (tchaton)")
 @pytest.mark.skipif(not _is_docker_available(), reason="docker is required for this test.")
 @pytest.mark.skipif(not check_if_redis_running(), reason="redis is required for this test.")
 @_RunIf(skip_windows=True)

tests/tests_lite/test_parity.py

@@ -162,7 +162,7 @@ def run(rank, model, train_dataloader, num_epochs, precision, accelerator, tmpdi
         _atomic_save(model.state_dict(), os.path.join(tmpdir, "model_spawn.pt"))
 
 
-@pytest.mark.skipif(True, reason="Skipping as it takes 80 seconds.")
+@pytest.mark.skip(reason="Skipping as it takes 80 seconds.")
 @RunIf(min_cuda_gpus=2)
 @pytest.mark.parametrize(
     "precision, strategy, devices, accelerator",