Prepare for 2024.5.0 release (#1337)

* Prepare for 2024.5.0 release

* Override style check's default Rust toolchain

* Drop mlflow from GPU CI deps

* Add python 3.11 GPU CI job

* Bump docker-py to 7.1.0 to get around requests breakage

* Don't check dtype for scalar timestamps GPU tests

.github/workflows/style.yml

@@ -15,6 +15,11 @@ jobs:
     steps:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
+      - uses: actions-rs/toolchain@v1
+        with:
+            toolchain: 1.72
+            components: clippy
+            default: true
       - uses: actions-rs/toolchain@v1
         with:
             toolchain: nightly

Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "dask-sql"
 repository = "https://github.com/dask-contrib/dask-sql"
-version = "2024.3.0"
+version = "2024.5.0"
 description = "Bindings for DataFusion used by Dask-SQL"
 readme = "README.md"
 license = "Apache-2.0"

continuous_integration/environment-3.10.yaml

@@ -5,14 +5,15 @@ dependencies:
 - c-compiler
 - dask>=2024.4.1
 - dask-expr>=1.0.11
+- docker-py>=7.1.0
 - fastapi>=0.92.0
 - fugue>=0.7.3
 - httpx>=0.24.1
 - intake>=0.6.0
 - jsonschema
 - lightgbm
 - maturin>=1.3,<1.4
-- mlflow>=2.9
+- mlflow>=2.10
 - mock
 - numpy>=1.22.4
 - pandas>=2

continuous_integration/environment-3.11.yaml

@@ -5,14 +5,15 @@ dependencies:
 - c-compiler
 - dask>=2024.4.1
 - dask-expr>=1.0.11
+- docker-py>=7.1.0
 - fastapi>=0.92.0
 - fugue>=0.7.3
 - httpx>=0.24.1
 - intake>=0.6.0
 - jsonschema
 - lightgbm
 - maturin>=1.3,<1.4
-- mlflow>=2.9
+- mlflow>=2.10
 - mock
 - numpy>=1.22.4
 - pandas>=2

continuous_integration/environment-3.12.yaml

@@ -5,6 +5,7 @@ dependencies:
 - c-compiler
 - dask>=2024.4.1
 - dask-expr>=1.0.11
+- docker-py>=7.1.0
 - fastapi>=0.92.0
 - fugue>=0.7.3
 - httpx>=0.24.1
@@ -13,7 +14,7 @@ dependencies:
 - lightgbm
 - maturin>=1.3,<1.4
 # TODO: add once mlflow 3.12 builds are available
-# - mlflow>=2.9
+# - mlflow>=2.10
 - mock
 - numpy>=1.22.4
 - pandas>=2

continuous_integration/environment-3.9.yaml

@@ -5,14 +5,15 @@ dependencies:
 - c-compiler
 - dask=2024.4.1
 - dask-expr=1.0.11
+- docker-py>=7.1.0
 - fastapi=0.92.0
 - fugue=0.7.3
 - httpx=0.24.1
 - intake=0.6.0
 - jsonschema
 - lightgbm
 - maturin=1.3
-- mlflow=2.9
+- mlflow=2.10
 - mock
 - numpy=1.22.4
 - pandas=2

continuous_integration/gpuci/axis.yaml

@@ -1,6 +1,7 @@
 PYTHON_VER:
 - "3.9"
 - "3.10"
+- "3.11"
 
 CUDA_VER:
 - "11.8.0"

continuous_integration/gpuci/environment-3.10.yaml

@@ -18,7 +18,6 @@ dependencies:
 - jsonschema
 - lightgbm
 - maturin>=1.3,<1.4
-- mlflow>=2.9
 - mock
 - numpy>=1.22.4
 - pandas>=2

continuous_integration/gpuci/environment-3.11.yaml

@@ -0,0 +1,53 @@
+name: dask-sql
+channels:
+- rapidsai
+- rapidsai-nightly
+- dask/label/dev
+- conda-forge
+- nvidia
+- nodefaults
+dependencies:
+- c-compiler
+- zlib
+- dask>=2024.4.1
+- dask-expr>=1.0.11
+- fastapi>=0.92.0
+- fugue>=0.7.3
+- httpx>=0.24.1
+- intake>=0.6.0
+- jsonschema
+- lightgbm
+- maturin>=1.3,<1.4
+- mock
+- numpy>=1.22.4
+- pandas>=2
+- pre-commit
+- prompt_toolkit>=3.0.8
+- psycopg2
+- pyarrow>=14.0.1
+- pygments>=2.7.1
+- pyhive
+- pytest-cov
+- pytest-rerunfailures
+- pytest-xdist
+- pytest
+- python=3.11
+- py-xgboost>=2.0.3
+- scikit-learn>=1.0.0
+- sphinx
+- sqlalchemy
+- tpot>=0.12.0
+# FIXME: https://github.com/fugue-project/fugue/issues/526
+- triad<0.9.2
+- tzlocal>=2.1
+- uvicorn>=0.14
+# GPU-specific requirements
+- cudatoolkit=11.8
+- cudf=24.06
+- cuml=24.06
+- dask-cudf=24.06
+- dask-cuda=24.06
+- ucx-proc=*=gpu
+- ucx-py=0.38
+- xgboost=*=rapidsai_py*
+- libxgboost=*=rapidsai_h*

continuous_integration/gpuci/environment-3.9.yaml

@@ -18,7 +18,6 @@ dependencies:
 - jsonschema
 - lightgbm
 - maturin>=1.3,<1.4
-- mlflow>=2.9
 - mock
 - numpy>=1.22.4
 - pandas>=2

tests/integration/test_rex.py

@@ -1144,7 +1144,7 @@ def test_scalar_timestamps(c, gpu):
         }
     )
     df1 = c.sql("SELECT CEIL(to_timestamp(d) TO DAY) AS ceil_to_day FROM df")
-    assert_eq(df1, expected_df)
+    assert_eq(df1, expected_df, check_dtype=(not gpu))
     df2 = c.sql("SELECT CEIL(CAST(d AS TIMESTAMP) TO DAY) AS ceil_to_day FROM df")
     assert_eq(df2, expected_df)
 
@@ -1154,7 +1154,7 @@ def test_scalar_timestamps(c, gpu):
         }
     )
     df1 = c.sql("SELECT FLOOR(to_timestamp(d) TO DAY) AS floor_to_day FROM df")
-    assert_eq(df1, expected_df)
+    assert_eq(df1, expected_df, check_dtype=(not gpu))
     df2 = c.sql("SELECT FLOOR(CAST(d AS TIMESTAMP) TO DAY) AS floor_to_day FROM df")
     assert_eq(df2, expected_df)