Bump to nvcomp 3.0.6. (rapidsai#15128)

This PR bumps nvcomp to 3.0.6. This is needed as a hotfix for rapidsai#15096.

Depends on:
- conda-forge/nvcomp-feedstock#14
- rapidsai/rapids-cmake#542
- rapidsai/kvikio#346

Authors:
   - Bradley Dice (https://github.com/bdice)

Approvers:
   - Lawrence Mitchell (https://github.com/wence-)
   - Ray Douglass (https://github.com/raydouglass)

conda/environments/all_cuda-118_arch-x86_64.yaml

@@ -62,7 +62,7 @@ dependencies:
 - numpy>=1.21,<1.25
 - numpydoc
 - nvcc_linux-64=11.8
-- nvcomp==3.0.5
+- nvcomp==3.0.6
 - nvtx>=0.2.1
 - packaging
 - pandas>=1.3,<1.6.0dev0

conda/environments/all_cuda-120_arch-x86_64.yaml

@@ -60,7 +60,7 @@ dependencies:
 - numba>=0.57
 - numpy>=1.21,<1.25
 - numpydoc
-- nvcomp==3.0.5
+- nvcomp==3.0.6
 - nvtx>=0.2.1
 - packaging
 - pandas>=1.3,<1.6.0dev0

conda/recipes/libcudf/conda_build_config.yaml

@@ -38,7 +38,7 @@ spdlog_version:
   - ">=1.12.0,<1.13"
 
 nvcomp_version:
-  - "=3.0.5"
+  - "=3.0.6"
 
 zlib_version:
   - ">=1.2.13"

dependencies.yaml

@@ -251,7 +251,7 @@ dependencies:
           - libkvikio==24.2.*
           - librdkafka>=1.9.0,<1.10.0a0
           # Align nvcomp version with rapids-cmake
-          - nvcomp==3.0.5
+          - nvcomp==3.0.6
           - spdlog>=1.12.0,<1.13
   build_wheels:
     common:

python/cudf/cudf/tests/test_parquet.py

@@ -3040,3 +3040,14 @@ def test_parquet_reader_multiindex():
 def test_parquet_reader_engine_error():
     with pytest.raises(ValueError):
         cudf.read_parquet(BytesIO(), engine="abc")
+
+
+def test_parquet_reader_zstd_huff_tables(datadir):
+    # Ensure that this zstd-compressed file does not overrun buffers. The
+    # problem was fixed in nvcomp 3.0.6.
+    # See https://github.com/rapidsai/cudf/issues/15096
+    fname = datadir / "zstd_huff_tables_bug.parquet"
+
+    expected = pa.parquet.read_table(fname).to_pandas()
+    actual = cudf.read_parquet(fname)
+    assert_eq(actual, expected)