fix up tests

rapidsai · Nov 15, 2024 · 3133ee1 · 3133ee1
1 parent 097decb
commit 3133ee1
Show file tree

Hide file tree

Showing 7 changed files with 18 additions and 23 deletions.
diff --git a/python/pylibcudf/pylibcudf/contiguous_split.pxd b/python/pylibcudf/pylibcudf/contiguous_split.pxd
@@ -1,12 +1,25 @@
 # Copyright (c) 2024, NVIDIA CORPORATION.
 
+from libc.stdint cimport uint8_t
 from libcpp.memory cimport unique_ptr
+from libcpp.vector cimport vector
 from pylibcudf.libcudf.contiguous_split cimport packed_columns
 
 from .gpumemoryview cimport gpumemoryview
 from .table cimport Table
 
 
+cdef class HostBuffer:
+    cdef unique_ptr[vector[uint8_t]] c_obj
+    cdef size_t nbytes
+    cdef Py_ssize_t[1] shape
+    cdef Py_ssize_t[1] strides
+
+    @staticmethod
+    cdef HostBuffer from_unique_ptr(
+        unique_ptr[vector[uint8_t]] vec
+    )
+
 cdef class PackedColumns:
     cdef unique_ptr[packed_columns] c_obj
 

diff --git a/python/pylibcudf/pylibcudf/contiguous_split.pyx b/python/pylibcudf/pylibcudf/contiguous_split.pyx
@@ -29,10 +29,6 @@ __all__ = [
 
 cdef class HostBuffer:
     """Owning host buffer that implements the buffer protocol"""
-    cdef unique_ptr[vector[uint8_t]] c_obj
-    cdef size_t nbytes
-    cdef Py_ssize_t[1] shape
-    cdef Py_ssize_t[1] strides
 
     @staticmethod
     cdef HostBuffer from_unique_ptr(

diff --git a/python/pylibcudf/pylibcudf/io/parquet.pxd b/python/pylibcudf/pylibcudf/io/parquet.pxd
@@ -50,9 +50,6 @@ cpdef read_parquet(
 cdef class ParquetWriterOptions:
     cdef parquet_writer_options c_obj
 
-    @staticmethod
-    cdef ParquetWriterOptionsBuilder builder(SinkInfo sink, Table table)
-
     cpdef void set_partitions(self, list partitions)
 
     cpdef void set_column_chunks_file_paths(self, list file_paths)

diff --git a/python/pylibcudf/pylibcudf/io/parquet.pyx b/python/pylibcudf/pylibcudf/io/parquet.pyx
@@ -249,7 +249,7 @@ cpdef read_parquet(
 cdef class ParquetWriterOptions:
 
     @staticmethod
-    cdef ParquetWriterOptionsBuilder builder(SinkInfo sink, Table table):
+    def builder(SinkInfo sink, Table table):
         """
         Create builder to create ParquetWriterOptionsBuilder.
 
@@ -287,9 +287,9 @@ cdef class ParquetWriterOptions:
         cdef vector[partition_info] c_partions
         cdef PartitionInfo partition
 
-        c_obj.reserve(len(partitions))
+        c_partions.reserve(len(partitions))
         for partition in partitions:
-            c_obj.push_back(partition.c_obj)
+            c_partions.push_back(partition.c_obj)
 
         self.c_obj.set_partitions(c_partions)
 

diff --git a/python/pylibcudf/pylibcudf/io/types.pxd b/python/pylibcudf/pylibcudf/io/types.pxd
@@ -25,9 +25,6 @@ from pylibcudf.table cimport Table
 cdef class PartitionInfo:
     cdef partition_info c_obj
 
-    @staticmethod
-    cdef PartitionInfo from_start_and_num(int start_row, int num_rows)
-
 cdef class ColumnInMetadata:
     cdef column_in_metadata c_obj
 
@@ -59,7 +56,6 @@ cdef class ColumnInMetadata:
     cpdef str get_name(self)
 
 cdef class TableInputMetadata:
-    cdef public Table table
     cdef table_input_metadata c_obj
 
 cdef class TableWithMetadata:

diff --git a/python/pylibcudf/pylibcudf/io/types.pyx b/python/pylibcudf/pylibcudf/io/types.pyx
@@ -61,7 +61,7 @@ cdef class PartitionInfo:
     num_rows : int
         The number of rows in the partition.
     """
-    def __init__(int start_row, int num_rows):
+    def __init__(self, int start_row, int num_rows):
         self.c_obj = partition_info(start_row, num_rows)
 
 
@@ -265,10 +265,6 @@ cdef class TableInputMetadata:
     """
     def __init__(self, Table table):
         self.c_obj = table_input_metadata(table.view())
-        self.column_metadata = [
-            ColumnInMetadata.from_metadata(metadata)
-            for metadata in self.c_obj.column_metadata
-        ]
 
 
 cdef class TableWithMetadata:

diff --git a/python/pylibcudf/pylibcudf/tests/io/test_parquet.py b/python/pylibcudf/pylibcudf/tests/io/test_parquet.py
@@ -125,7 +125,6 @@ def test_read_parquet_filters(
         plc.io.types.StatisticsFreq.STATISTICS_COLUMN,
     ],
 )
-@pytest.mark.parametrize("int96_timestamps", [True, False])
 @pytest.mark.parametrize("write_v2_headers", [True, False])
 @pytest.mark.parametrize(
     "dictionary_policy",
@@ -150,7 +149,6 @@ def test_write_parquet(
     table_data,
     compression,
     stats_level,
-    int96_timestamps,
     write_v2_headers,
     dictionary_policy,
     utc_timestamps,
@@ -174,7 +172,6 @@ def test_write_parquet(
         .key_value_metadata(user_data)
         .compression(compression)
         .stats_level(stats_level)
-        .int96_timestamps(int96_timestamps)
         .write_v2_headers(write_v2_headers)
         .dictionary_policy(dictionary_policy)
         .utc_timestamps(utc_timestamps)
@@ -197,4 +194,4 @@ def test_write_parquet(
         options.set_max_dictionary_size(max_dictionary_size)
 
     result = plc.io.parquet.write_parquet(options)
-    assert isinstance(result, plc.io.parquet.BufferArrayFromVector)
+    assert isinstance(result, plc.contiguous_split.HostBuffer)