Enforce input data is not object. (#6927)

* Check for object data type.
* Allow strided arrays with greater underlying buffer size.

python-package/xgboost/core.py

@@ -233,6 +233,9 @@ def _numpy2ctypes_type(dtype):
 
 
 def _array_interface(data: np.ndarray) -> bytes:
+    assert (
+        data.dtype.hasobject is False
+    ), "Input data contains `object` dtype.  Expecting numeric data."
     interface = data.__array_interface__
     if "mask" in interface:
         interface["mask"] = interface["mask"].__array_interface__
@@ -1908,8 +1911,8 @@ class Booster(object):
                 )
 
         if isinstance(data, np.ndarray):
-            from .data import _maybe_np_slice
-            data = _maybe_np_slice(data, data.dtype)
+            from .data import _ensure_np_dtype
+            data, _ = _ensure_np_dtype(data, data.dtype)
             _check_call(
                 _LIB.XGBoosterPredictFromDense(
                     self.handle,

python-package/xgboost/data.py

@@ -104,6 +104,13 @@ def _is_numpy_array(data):
     return isinstance(data, (np.ndarray, np.matrix))
 
 
+def _ensure_np_dtype(data, dtype):
+    if data.dtype.hasobject:
+        data = data.astype(np.float32, copy=False)
+        dtype = np.float32
+    return data, dtype
+
+
 def _maybe_np_slice(data, dtype):
     '''Handle numpy slice.  This can be removed if we use __array_interface__.
     '''
@@ -118,6 +125,7 @@ def _maybe_np_slice(data, dtype):
             data = np.array(data, copy=False, dtype=dtype)
     except AttributeError:
         data = np.array(data, copy=False, dtype=dtype)
+    data, dtype = _ensure_np_dtype(data, dtype)
     return data