TST: Remove tm.rands/rands_array (#54368)

* remove tm.rands

* remove rands array

* Address failures

* Use unique values

Author: mroeschke

asv_bench/benchmarks/array.py

@@ -2,8 +2,6 @@
 
 import pandas as pd
 
-from .pandas_vb_common import tm
-
 
 class BooleanArray:
     def setup(self):
@@ -56,7 +54,7 @@ def time_from_tuples(self):
 class StringArray:
     def setup(self):
         N = 100_000
-        values = tm.rands_array(3, N)
+        values = np.array([str(i) for i in range(N)], dtype=object)
         self.values_obj = np.array(values, dtype="object")
         self.values_str = np.array(values, dtype="U")
         self.values_list = values.tolist()
@@ -80,7 +78,7 @@ def setup(self, multiple_chunks):
             import pyarrow as pa
         except ImportError:
             raise NotImplementedError
-        strings = tm.rands_array(3, 10_000)
+        strings = np.array([str(i) for i in range(10_000)], dtype=object)
         if multiple_chunks:
             chunks = [strings[i : i + 100] for i in range(0, len(strings), 100)]
             self.array = pd.arrays.ArrowStringArray(pa.chunked_array(chunks))
@@ -127,7 +125,7 @@ def setup(self, dtype, hasna):
         elif dtype == "int64[pyarrow]":
             data = np.arange(N)
         elif dtype == "string[pyarrow]":
-            data = tm.rands_array(10, N)
+            data = np.array([str(i) for i in range(N)], dtype=object)
         elif dtype == "timestamp[ns][pyarrow]":
             data = pd.date_range("2000-01-01", freq="s", periods=N)
         else:

asv_bench/benchmarks/series_methods.py

@@ -104,7 +104,7 @@ def setup(self, dtype):
             data = np.arange(N)
             na_value = NA
         elif dtype in ("string", "string[pyarrow]"):
-            data = tm.rands_array(5, N)
+            data = np.array([str(i) * 5 for i in range(N)], dtype=object)
             na_value = NA
         else:
             raise NotImplementedError

asv_bench/benchmarks/strings.py

@@ -34,8 +34,8 @@ class Construction:
     dtype_mapping = {"str": "str", "string[python]": object, "string[pyarrow]": object}
 
     def setup(self, pd_type, dtype):
-        series_arr = tm.rands_array(
-            nchars=10, size=10**5, dtype=self.dtype_mapping[dtype]
+        series_arr = np.array(
+            [str(i) * 10 for i in range(100_000)], dtype=self.dtype_mapping[dtype]
         )
         if pd_type == "series":
             self.arr = series_arr
@@ -276,7 +276,7 @@ def time_iter(self, dtype):
 
 class StringArrayConstruction:
     def setup(self):
-        self.series_arr = tm.rands_array(nchars=10, size=10**5)
+        self.series_arr = np.array([str(i) * 10 for i in range(10**5)], dtype=object)
         self.series_arr_nan = np.concatenate([self.series_arr, np.array([NA] * 1000)])
 
     def time_string_array_construction(self):

pandas/_testing/__init__.py

@@ -54,10 +54,6 @@
     round_trip_pickle,
     write_to_compressed,
 )
-from pandas._testing._random import (
-    rands,
-    rands_array,
-)
 from pandas._testing._warnings import (
     assert_produces_warning,
     maybe_produces_warning,
@@ -349,6 +345,22 @@ def to_array(obj):
 # Others
 
 
+def rands_array(
+    nchars, size: int, dtype: NpDtype = "O", replace: bool = True
+) -> np.ndarray:
+    """
+    Generate an array of byte strings.
+    """
+    chars = np.array(list(string.ascii_letters + string.digits), dtype=(np.str_, 1))
+    retval = (
+        np.random.default_rng(2)
+        .choice(chars, size=nchars * np.prod(size), replace=replace)
+        .view((np.str_, nchars))
+        .reshape(size)
+    )
+    return retval.astype(dtype)
+
+
 def getCols(k) -> str:
     return string.ascii_uppercase[:k]
 
@@ -1127,7 +1139,6 @@ def shares_memory(left, right) -> bool:
     "NULL_OBJECTS",
     "OBJECT_DTYPES",
     "raise_assert_detail",
-    "rands",
     "reset_display_options",
     "raises_chained_assignment_error",
     "round_trip_localpath",

pandas/_testing/_io.py

@@ -9,6 +9,7 @@
     Any,
     Callable,
 )
+import uuid
 import zipfile
 
 from pandas.compat import (
@@ -18,7 +19,6 @@
 from pandas.compat._optional import import_optional_dependency
 
 import pandas as pd
-from pandas._testing._random import rands
 from pandas._testing.contexts import ensure_clean
 
 if TYPE_CHECKING:
@@ -56,7 +56,7 @@ def round_trip_pickle(
     """
     _path = path
     if _path is None:
-        _path = f"__{rands(10)}__.pickle"
+        _path = f"__{uuid.uuid4()}__.pickle"
     with ensure_clean(_path) as temp_path:
         pd.to_pickle(obj, temp_path)
         return pd.read_pickle(temp_path)

pandas/_testing/_random.py

@@ -881,7 +881,7 @@ def test_add_frames(self, first, second, expected):
     # TODO: This came from series.test.test_operators, needs cleanup
     def test_series_frame_radd_bug(self, fixed_now_ts):
         # GH#353
-        vals = Series(tm.rands_array(5, 10))
+        vals = Series(tm.makeStringIndex())
         result = "foo_" + vals
         expected = vals.map(lambda x: "foo_" + x)
         tm.assert_series_equal(result, expected)

pandas/tests/arithmetic/test_numeric.py

@@ -272,7 +272,7 @@ def test_getitem_series_integer_with_missing_raises(self, data, idx):
         msg = "Cannot index with an integer indexer containing NA values"
         # TODO: this raises KeyError about labels not found (it tries label-based)
 
-        ser = pd.Series(data, index=[tm.rands(4) for _ in range(len(data))])
+        ser = pd.Series(data, index=[chr(100 + i) for i in range(len(data))])
         with pytest.raises(ValueError, match=msg):
             ser[idx]
 

pandas/tests/extension/base/getitem.py

@@ -197,7 +197,7 @@ def test_setitem_integer_with_missing_raises(self, data, idx, box_in_series):
         # TODO(xfail) this raises KeyError about labels not found (it tries label-based)
         # for list of labels with Series
         if box_in_series:
-            arr = pd.Series(data, index=[tm.rands(4) for _ in range(len(data))])
+            arr = pd.Series(data, index=[chr(100 + i) for i in range(len(data))])
 
         msg = "Cannot index with an integer indexer containing NA values"
         with pytest.raises(ValueError, match=msg):

pandas/tests/extension/base/setitem.py

@@ -203,7 +203,7 @@ def test_timestamp_compare(self, left, right):
                 "dates2": pd.date_range("20010102", periods=10),
                 "intcol": np.random.default_rng(2).integers(1000000000, size=10),
                 "floatcol": np.random.default_rng(2).standard_normal(10),
-                "stringcol": list(tm.rands(10)),
+                "stringcol": [chr(100 + i) for i in range(10)],
             }
         )
         df.loc[np.random.default_rng(2).random(len(df)) > 0.5, "dates2"] = pd.NaT