Move tune imports to new locations (#253)

* Move tune imports to new locations Signed-off-by: Antoni Baum <[email protected]> * Fix master test Signed-off-by: Antoni Baum <[email protected]> * Bump Ray version Signed-off-by: Antoni Baum <[email protected]> * Remove debug Signed-off-by: Antoni Baum <[email protected]> Signed-off-by: Antoni Baum <[email protected]>
ray-project · Oct 5, 2022 · 0ddd93a · 0ddd93a
1 parent 0e1bd33
commit 0ddd93a
Show file tree

Hide file tree

Showing 7 changed files with 21 additions and 19 deletions.
diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
@@ -18,7 +18,7 @@ jobs:
         python -m pip install --upgrade pip
         python -m pip install -U pytest
         python -m pip install codecov
-        python -m pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-2.0.0.dev0-cp37-cp37m-manylinux2014_x86_64.whl
+        python -m pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-3.0.0.dev0-cp37-cp37m-manylinux2014_x86_64.whl
         python -m pip install -U -q scikit-learn scikit-optimize hyperopt hpbandster ConfigSpace scipy dataclasses optuna keras
         if [ -f requirements-test.txt ]; then python -m pip install -r requirements-test.txt; fi
     - name: Install package

diff --git a/examples/custom_searcher_example.py b/examples/custom_searcher_example.py
@@ -5,7 +5,7 @@
 
 from tune_sklearn import TuneSearchCV
 from ray import tune
-from ray.tune.suggest.hebo import HEBOSearch
+from ray.tune.search.hebo import HEBOSearch
 from sklearn.ensemble import RandomForestClassifier
 from sklearn import datasets
 from sklearn.model_selection import train_test_split
@@ -28,7 +28,7 @@
 searcher = HEBOSearch()
 
 # It is also possible to use user-defined Searchers, as long as
-# they inherit from ray.tune.suggest.Searcher and have the following
+# they inherit from ray.tune.search.Searcher and have the following
 # attributes: _space, _metric, _mode
 
 tune_search = TuneSearchCV(

diff --git a/setup.py b/setup.py
@@ -27,4 +27,6 @@
         os.path.join(ROOT_DIR, "README.md"), "r", encoding="utf-8").read(),
     long_description_content_type="text/markdown",
     url="https://github.com/ray-project/tune-sklearn",
-    install_requires=["scikit-learn", "scipy", "ray[tune]", "numpy>=1.16"])
+    install_requires=[
+        "scikit-learn", "scipy", "ray[tune]>=2.0.0", "numpy>=1.16"
+    ])
diff --git a/tests/test_randomizedsearch.py b/tests/test_randomizedsearch.py
@@ -692,7 +692,7 @@ def testOptuna(self):
         self._test_method("optuna")
 
     def testCustomSearcher(self):
-        from ray.tune.suggest.hyperopt import HyperOptSearch
+        from ray.tune.search.hyperopt import HyperOptSearch
 
         class CustomSearcher(HyperOptSearch):
             pass
@@ -708,7 +708,7 @@ class ThisShouldRaiseAnExc:
         self._test_method(CustomSearcher())
 
     def testCustomSearcherWithSearchSpaceException(self):
-        from ray.tune.suggest.hyperopt import HyperOptSearch
+        from ray.tune.search.hyperopt import HyperOptSearch
         from hyperopt import hp
 
         class CustomSearcher(HyperOptSearch):
@@ -737,7 +737,7 @@ class CustomSearcher(HyperOptSearch):
             " its mode") in str(exc.exception))
 
     def testCustomSearcherWithSearchSpace(self):
-        from ray.tune.suggest.hyperopt import HyperOptSearch
+        from ray.tune.search.hyperopt import HyperOptSearch
         from hyperopt import hp
 
         class CustomSearcher(HyperOptSearch):
@@ -827,7 +827,7 @@ def testBayesianPointsToEvaluate(self):
         self._test_points_to_evaluate("bayesian")
 
     def testHyperoptPointsToEvaluate(self):
-        from ray.tune.suggest.hyperopt import HyperOptSearch
+        from ray.tune.search.hyperopt import HyperOptSearch
         # Skip test if category conversion is not available
         if not hasattr(HyperOptSearch, "_convert_categories_to_indices"):
             self.skipTest(f"The current version of Ray does not support the "

diff --git a/tune_sklearn/list_searcher.py b/tune_sklearn/list_searcher.py
@@ -4,7 +4,7 @@
 """
 
 from sklearn.model_selection import ParameterGrid
-from ray.tune.suggest.suggestion import Searcher
+from ray.tune.search.searcher import Searcher
 import random
 
 

diff --git a/tune_sklearn/tune_basesearch.py b/tune_sklearn/tune_basesearch.py
@@ -30,7 +30,7 @@
 import numbers
 
 import ray
-from ray.tune.trial import Trial
+from ray.tune.experiment.trial import Trial
 from ray.tune.schedulers import (
     PopulationBasedTraining, AsyncHyperBandScheduler, HyperBandScheduler,
     MedianStoppingRule, TrialScheduler, ASHAScheduler, HyperBandForBOHB)

diff --git a/tune_sklearn/tune_search.py b/tune_sklearn/tune_search.py
@@ -9,15 +9,15 @@
 from sklearn.base import clone
 
 from ray import tune
-from ray.tune.sample import Domain
-from ray.tune.suggest import (ConcurrencyLimiter, BasicVariantGenerator,
-                              Searcher)
-from ray.tune.suggest.bohb import TuneBOHB
+from ray.tune.search.sample import Domain
+from ray.tune.search import (ConcurrencyLimiter, BasicVariantGenerator,
+                             Searcher)
+from ray.tune.search.bohb import TuneBOHB
 from ray.tune.schedulers import HyperBandForBOHB
 from ray.tune.stopper import CombinedStopper
-from ray.tune.suggest.skopt import SkOptSearch
-from ray.tune.suggest.hyperopt import HyperOptSearch
-from ray.tune.suggest.optuna import OptunaSearch
+from ray.tune.search.skopt import SkOptSearch
+from ray.tune.search.hyperopt import HyperOptSearch
+from ray.tune.search.optuna import OptunaSearch
 
 from tune_sklearn.utils import check_is_pipeline, MaximumIterationStopper
 from tune_sklearn.tune_basesearch import TuneBaseSearchCV
@@ -254,7 +254,7 @@ class TuneSearchCV(TuneBaseSearchCV):
             resource_param (max_iter or n_estimators) is
             incremented by `max resource value // max_iters`.
         search_optimization ("random" or "bayesian" or "bohb" or "hyperopt"
-            or "optuna" or `ray.tune.suggest.Searcher` instance):
+            or "optuna" or `ray.tune.search.Searcher` instance):
             Randomized search is invoked with ``search_optimization`` set to
             ``"random"`` and behaves like scikit-learn's
             ``RandomizedSearchCV``.
@@ -350,7 +350,7 @@ def __init__(self,
             raise ValueError(
                 "Search optimization must be one of "
                 f"{', '.join(list(available_optimizations.values()))} "
-                "or a ray.tune.suggest.Searcher instance.")
+                "or a ray.tune.search.Searcher instance.")
 
         if isinstance(self._search_optimization_lower, Searcher):
             if not hasattr(self._search_optimization_lower,