Disallow use of Pickle since numpy versions diverge

frameworks/AutoGluon/__init__.py

@@ -1,4 +1,3 @@
-
 from amlb.utils import call_script_in_same_dir
 from amlb.benchmark import TaskConfig
 from amlb.data import Dataset, DatasetType
@@ -8,8 +7,8 @@
 def setup(*args, **kwargs):
     call_script_in_same_dir(__file__, "setup.sh", *args, **kwargs)
 
-def run(dataset: Dataset, config: TaskConfig):
 
+def run(dataset: Dataset, config: TaskConfig):
     if dataset.type == DatasetType.timeseries:
         return run_autogluon_timeseries(dataset, config)
     else:
@@ -18,23 +17,33 @@ def run(dataset: Dataset, config: TaskConfig):
 
 def run_autogluon_tabular(dataset: Dataset, config: TaskConfig):
     from frameworks.shared.caller import run_in_venv
+
     data = dict(
-        train=dict(path=dataset.train.data_path('parquet')),
-        test=dict(path=dataset.test.data_path('parquet')),
-        target=dict(
-            name=dataset.target.name,
-            classes=dataset.target.values
-        ),
+        train=dict(path=dataset.train.data_path("parquet")),
+        test=dict(path=dataset.test.data_path("parquet")),
+        target=dict(name=dataset.target.name, classes=dataset.target.values),
         problem_type=dataset.type.name,  # AutoGluon problem_type is using same names as amlb.data.DatasetType
     )
     if config.measure_inference_time:
-        data["inference_subsample_files"] = dataset.inference_subsample_files(fmt="parquet")
+        data["inference_subsample_files"] = dataset.inference_subsample_files(
+            fmt="parquet"
+        )
+
+    options = {"serialization": {"numpy_allow_pickle": False}}
+
+    return run_in_venv(
+        __file__,
+        "exec.py",
+        input_data=data,
+        dataset=dataset,
+        config=config,
+        options=options,
+    )
 
-    return run_in_venv(__file__, "exec.py",
-                       input_data=data, dataset=dataset, config=config)
 
 def run_autogluon_timeseries(dataset: Dataset, config: TaskConfig):
     from frameworks.shared.caller import run_in_venv
+
     dataset = deepcopy(dataset)
 
     data = dict(
@@ -50,5 +59,6 @@ def run_autogluon_timeseries(dataset: Dataset, config: TaskConfig):
         repeated_item_id=dataset.repeated_item_id,
     )
 
-    return run_in_venv(__file__, "exec_ts.py",
-                       input_data=data, dataset=dataset, config=config)
+    return run_in_venv(
+        __file__, "exec_ts.py", input_data=data, dataset=dataset, config=config
+    )