Merge branch 'branch-24.12' into pandas_upgrade

cpp/include/cudf/detail/utilities/logger.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2023, NVIDIA CORPORATION.
+ * Copyright (c) 2023-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -19,9 +19,9 @@
 #include <cudf/utilities/logger.hpp>
 
 // Log messages that require computation should only be used at level TRACE and DEBUG
-#define CUDF_LOG_TRACE(...)    SPDLOG_LOGGER_TRACE(&cudf::logger(), __VA_ARGS__)
-#define CUDF_LOG_DEBUG(...)    SPDLOG_LOGGER_DEBUG(&cudf::logger(), __VA_ARGS__)
-#define CUDF_LOG_INFO(...)     SPDLOG_LOGGER_INFO(&cudf::logger(), __VA_ARGS__)
-#define CUDF_LOG_WARN(...)     SPDLOG_LOGGER_WARN(&cudf::logger(), __VA_ARGS__)
-#define CUDF_LOG_ERROR(...)    SPDLOG_LOGGER_ERROR(&cudf::logger(), __VA_ARGS__)
-#define CUDF_LOG_CRITICAL(...) SPDLOG_LOGGER_CRITICAL(&cudf::logger(), __VA_ARGS__)
+#define CUDF_LOG_TRACE(...)    SPDLOG_LOGGER_TRACE(&cudf::detail::logger(), __VA_ARGS__)
+#define CUDF_LOG_DEBUG(...)    SPDLOG_LOGGER_DEBUG(&cudf::detail::logger(), __VA_ARGS__)
+#define CUDF_LOG_INFO(...)     SPDLOG_LOGGER_INFO(&cudf::detail::logger(), __VA_ARGS__)
+#define CUDF_LOG_WARN(...)     SPDLOG_LOGGER_WARN(&cudf::detail::logger(), __VA_ARGS__)
+#define CUDF_LOG_ERROR(...)    SPDLOG_LOGGER_ERROR(&cudf::detail::logger(), __VA_ARGS__)
+#define CUDF_LOG_CRITICAL(...) SPDLOG_LOGGER_CRITICAL(&cudf::detail::logger(), __VA_ARGS__)

cpp/include/cudf/utilities/logger.hpp

@@ -22,6 +22,10 @@
 
 namespace CUDF_EXPORT cudf {
 
+namespace detail {
+spdlog::logger& logger();
+}
+
 /**
  * @brief Returns the global logger.
  *
@@ -43,6 +47,8 @@ namespace CUDF_EXPORT cudf {
  *
  * @return spdlog::logger& The logger.
  */
-spdlog::logger& logger();
+[[deprecated(
+  "Support for direct access to spdlog loggers in cudf is planned for removal")]] spdlog::logger&
+logger();
 
 }  // namespace CUDF_EXPORT cudf

cpp/include/nvtext/edit_distance.hpp

@@ -57,7 +57,7 @@ namespace CUDF_EXPORT nvtext {
  * @param targets Strings to compute edit distance against `input`
  * @param stream CUDA stream used for device memory operations and kernel launches
  * @param mr Device memory resource used to allocate the returned column's device memory
- * @return New strings columns of with replaced strings
+ * @return New lists column of edit distance values
  */
 std::unique_ptr<cudf::column> edit_distance(
   cudf::strings_column_view const& input,

cpp/src/utilities/logger.cpp

@@ -74,8 +74,10 @@ struct logger_wrapper {
 
 }  // namespace
 
-spdlog::logger& cudf::logger()
+spdlog::logger& cudf::detail::logger()
 {
   static logger_wrapper wrapped{};
   return wrapped.logger_;
 }
+
+spdlog::logger& cudf::logger() { return cudf::detail::logger(); }

cpp/tests/utilities_tests/logger_tests.cpp

@@ -28,16 +28,17 @@ class LoggerTest : public cudf::test::BaseFixture {
   std::vector<spdlog::sink_ptr> prev_sinks;
 
  public:
-  LoggerTest() : prev_level{cudf::logger().level()}, prev_sinks{cudf::logger().sinks()}
+  LoggerTest()
+    : prev_level{cudf::detail::logger().level()}, prev_sinks{cudf::detail::logger().sinks()}
   {
-    cudf::logger().sinks() = {std::make_shared<spdlog::sinks::ostream_sink_mt>(oss)};
-    cudf::logger().set_formatter(
+    cudf::detail::logger().sinks() = {std::make_shared<spdlog::sinks::ostream_sink_mt>(oss)};
+    cudf::detail::logger().set_formatter(
       std::unique_ptr<spdlog::formatter>(new spdlog::pattern_formatter("%v")));
   }
   ~LoggerTest() override
   {
-    cudf::logger().set_level(prev_level);
-    cudf::logger().sinks() = prev_sinks;
+    cudf::detail::logger().set_level(prev_level);
+    cudf::detail::logger().sinks() = prev_sinks;
   }
 
   void clear_sink() { oss.str(""); }
@@ -46,32 +47,32 @@ class LoggerTest : public cudf::test::BaseFixture {
 
 TEST_F(LoggerTest, Basic)
 {
-  cudf::logger().critical("crit msg");
+  cudf::detail::logger().critical("crit msg");
   ASSERT_EQ(this->sink_content(), "crit msg\n");
 }
 
 TEST_F(LoggerTest, DefaultLevel)
 {
-  cudf::logger().trace("trace");
-  cudf::logger().debug("debug");
-  cudf::logger().info("info");
-  cudf::logger().warn("warn");
-  cudf::logger().error("error");
-  cudf::logger().critical("critical");
+  cudf::detail::logger().trace("trace");
+  cudf::detail::logger().debug("debug");
+  cudf::detail::logger().info("info");
+  cudf::detail::logger().warn("warn");
+  cudf::detail::logger().error("error");
+  cudf::detail::logger().critical("critical");
   ASSERT_EQ(this->sink_content(), "warn\nerror\ncritical\n");
 }
 
 TEST_F(LoggerTest, CustomLevel)
 {
-  cudf::logger().set_level(spdlog::level::warn);
-  cudf::logger().info("info");
-  cudf::logger().warn("warn");
+  cudf::detail::logger().set_level(spdlog::level::warn);
+  cudf::detail::logger().info("info");
+  cudf::detail::logger().warn("warn");
   ASSERT_EQ(this->sink_content(), "warn\n");
 
   this->clear_sink();
 
-  cudf::logger().set_level(spdlog::level::debug);
-  cudf::logger().trace("trace");
-  cudf::logger().debug("debug");
+  cudf::detail::logger().set_level(spdlog::level::debug);
+  cudf::detail::logger().trace("trace");
+  cudf::detail::logger().debug("debug");
   ASSERT_EQ(this->sink_content(), "debug\n");
 }

docs/cudf/source/user_guide/api_docs/pylibcudf/index.rst

@@ -49,3 +49,4 @@ This page provides API documentation for pylibcudf.
 
     io/index.rst
     strings/index.rst
+    nvtext/index.rst

docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/edit_distance.rst

@@ -0,0 +1,6 @@
+=============
+edit_distance
+=============
+
+.. automodule:: pylibcudf.nvtext.edit_distance
+   :members:

docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/index.rst

@@ -0,0 +1,7 @@
+nvtext
+======
+
+.. toctree::
+    :maxdepth: 1
+
+    edit_distance

python/cudf/cudf/_lib/nvtext/edit_distance.pyx

@@ -2,37 +2,23 @@
 
 from cudf.core.buffer import acquire_spill_lock
 
-from libcpp.memory cimport unique_ptr
-from libcpp.utility cimport move
-
-from pylibcudf.libcudf.column.column cimport column
-from pylibcudf.libcudf.column.column_view cimport column_view
-from pylibcudf.libcudf.nvtext.edit_distance cimport (
-    edit_distance as cpp_edit_distance,
-    edit_distance_matrix as cpp_edit_distance_matrix,
-)
+from pylibcudf cimport nvtext
 
 from cudf._lib.column cimport Column
 
 
 @acquire_spill_lock()
 def edit_distance(Column strings, Column targets):
-    cdef column_view c_strings = strings.view()
-    cdef column_view c_targets = targets.view()
-    cdef unique_ptr[column] c_result
-
-    with nogil:
-        c_result = move(cpp_edit_distance(c_strings, c_targets))
-
-    return Column.from_unique_ptr(move(c_result))
+    result = nvtext.edit_distance.edit_distance(
+        strings.to_pylibcudf(mode="read"),
+        targets.to_pylibcudf(mode="read")
+    )
+    return Column.from_pylibcudf(result)
 
 
 @acquire_spill_lock()
 def edit_distance_matrix(Column strings):
-    cdef column_view c_strings = strings.view()
-    cdef unique_ptr[column] c_result
-
-    with nogil:
-        c_result = move(cpp_edit_distance_matrix(c_strings))
-
-    return Column.from_unique_ptr(move(c_result))
+    result = nvtext.edit_distance.edit_distance_matrix(
+        strings.to_pylibcudf(mode="read")
+    )
+    return Column.from_pylibcudf(result)

python/cudf/cudf/_lib/string_casting.pyx

@@ -3,9 +3,6 @@
 from cudf._lib.column cimport Column
 
 from cudf._lib.scalar import as_device_scalar
-
-from cudf._lib.scalar cimport DeviceScalar
-
 from cudf._lib.types import SUPPORTED_NUMPY_TO_LIBCUDF_TYPES
 
 from libcpp.memory cimport unique_ptr
@@ -14,14 +11,6 @@ from libcpp.utility cimport move
 
 from pylibcudf.libcudf.column.column cimport column
 from pylibcudf.libcudf.column.column_view cimport column_view
-from pylibcudf.libcudf.scalar.scalar cimport string_scalar
-from pylibcudf.libcudf.strings.convert.convert_booleans cimport (
-    from_booleans as cpp_from_booleans,
-    to_booleans as cpp_to_booleans,
-)
-from pylibcudf.libcudf.strings.convert.convert_datetime cimport (
-    is_timestamp as cpp_is_timestamp,
-)
 from pylibcudf.libcudf.strings.convert.convert_floats cimport (
     from_floats as cpp_from_floats,
     to_floats as cpp_to_floats,
@@ -427,77 +416,21 @@ def stoul(Column input_col):
     return string_to_integer(input_col, cudf.dtype("uint64"))
 
 
-def _to_booleans(Column input_col, object string_true="True"):
-    """
-    Converting/Casting input column of type string to boolean column
-
-    Parameters
-    ----------
-    input_col : input column of type string
-    string_true : string that represents True
-
-    Returns
-    -------
-    A Column with string values cast to boolean
-    """
-
-    cdef DeviceScalar str_true = as_device_scalar(string_true)
-    cdef column_view input_column_view = input_col.view()
-    cdef const string_scalar* string_scalar_true = <const string_scalar*>(
-        str_true.get_raw_ptr())
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_to_booleans(
-                input_column_view,
-                string_scalar_true[0]))
-
-    return Column.from_unique_ptr(move(c_result))
-
-
 def to_booleans(Column input_col):
-
-    return _to_booleans(input_col)
-
-
-def _from_booleans(
-        Column input_col,
-        object string_true="True",
-        object string_false="False"):
-    """
-    Converting/Casting input column of type boolean to string column
-
-    Parameters
-    ----------
-    input_col : input column of type boolean
-    string_true : string that represents True
-    string_false : string that represents False
-
-    Returns
-    -------
-    A Column with boolean values cast to string
-    """
-
-    cdef DeviceScalar str_true = as_device_scalar(string_true)
-    cdef DeviceScalar str_false = as_device_scalar(string_false)
-    cdef column_view input_column_view = input_col.view()
-    cdef const string_scalar* string_scalar_true = <const string_scalar*>(
-        str_true.get_raw_ptr())
-    cdef const string_scalar* string_scalar_false = <const string_scalar*>(
-        str_false.get_raw_ptr())
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_from_booleans(
-                input_column_view,
-                string_scalar_true[0],
-                string_scalar_false[0]))
-
-    return Column.from_unique_ptr(move(c_result))
+    plc_column = plc.strings.convert.convert_booleans.to_booleans(
+        input_col.to_pylibcudf(mode="read"),
+        as_device_scalar("True").c_value,
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def from_booleans(Column input_col):
-    return _from_booleans(input_col)
+    plc_column = plc.strings.convert.convert_booleans.from_booleans(
+        input_col.to_pylibcudf(mode="read"),
+        as_device_scalar("True").c_value,
+        as_device_scalar("False").c_value,
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def int2timestamp(
@@ -520,11 +453,10 @@ def int2timestamp(
     A Column with date-time represented in string format
 
     """
-    cdef string c_timestamp_format = format.encode("UTF-8")
     return Column.from_pylibcudf(
         plc.strings.convert.convert_datetime.from_timestamps(
             input_col.to_pylibcudf(mode="read"),
-            c_timestamp_format,
+            format,
             names.to_pylibcudf(mode="read")
         )
     )
@@ -545,12 +477,11 @@ def timestamp2int(Column input_col, dtype, format):
 
     """
     dtype = dtype_to_pylibcudf_type(dtype)
-    cdef string c_timestamp_format = format.encode('UTF-8')
     return Column.from_pylibcudf(
         plc.strings.convert.convert_datetime.to_timestamps(
             input_col.to_pylibcudf(mode="read"),
             dtype,
-            c_timestamp_format
+            format
         )
     )
 
@@ -572,16 +503,11 @@ def istimestamp(Column input_col, str format):
     """
     if input_col.size == 0:
         return cudf.core.column.column_empty(0, dtype=cudf.dtype("bool"))
-    cdef column_view input_column_view = input_col.view()
-    cdef string c_timestamp_format = <string>str(format).encode('UTF-8')
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_is_timestamp(
-                input_column_view,
-                c_timestamp_format))
-
-    return Column.from_unique_ptr(move(c_result))
+    plc_column = plc.strings.convert.convert_datetime.is_timestamp(
+        input_col.to_pylibcudf(mode="read"),
+        format
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def timedelta2int(Column input_col, dtype, format):

python/cudf/cudf/core/reshape.py

@@ -681,7 +681,7 @@ def _tile(A, reps):
     nval = len(value_vars)
     dtype = min_unsigned_type(nval)
 
-    if not var_name:
+    if var_name is None:
         var_name = "variable"
 
     if not value_vars:

python/cudf/cudf/tests/test_reshape.py

@@ -119,6 +119,15 @@ def test_melt_str_scalar_id_var():
     assert_eq(result, expected)
 
 
+def test_melt_falsy_var_name():
+    df = cudf.DataFrame({"A": ["a", "b", "c"], "B": [1, 3, 5], "C": [2, 4, 6]})
+    result = cudf.melt(df, id_vars=["A"], value_vars=["B"], var_name="")
+    expected = pd.melt(
+        df.to_pandas(), id_vars=["A"], value_vars=["B"], var_name=""
+    )
+    assert_eq(result, expected)
+
+
 @pytest.mark.parametrize("num_cols", [1, 2, 10])
 @pytest.mark.parametrize("num_rows", [1, 2, 1000])
 @pytest.mark.parametrize(