search

BlazingDB / blazingsql

BlazingSQL is a lightweight, GPU accelerated, SQL engine for Python. Built on RAPIDS cuDF.
https://blazingsql.com
Apache License 2.0
1.92k stars 181 forks source link

[FEA] Support numpy>=1.20 #1432

Closed randerzander closed 3 years ago

randerzander commented 3 years ago

Related to cudf #7289.

BlazingSQL seems to require numpy=1.19 which breaks dask-cudf's describe functionality.

Is numpy 1.19 a hard requirement for BlazingSQL?

romulo-auccapuclla commented 3 years ago

@randerzander thanks for open this issue, regarding numpy=1.19 is not a hard requirement for BlazingSQL because its rely's on cudf libraries. Currently we are working to update some dependencies which will install numpy=1.20.2 in this PR https://github.com/BlazingDB/blazingsql/pull/1430, meanwhile you can build from source branch-0.19 with the following steps:

conda create --yes -n bsql python=3.8

conda activate bsql

conda install --yes -c conda-forge spdlog=1.7.0 google-cloud-cpp=1.25.0 ninja tqdm mysql-connector-cpp=8.0.23 libpq=13 sqlite=3 nlohmann_json

conda install --yes -c rapidsai-nightly -c nvidia -c conda-forge -c defaults dask-cuda=0.19 dask-cudf=0.19 ucx-py=0.19 ucx-proc=*=gpu cudf=0.19 python=3.8 cudatoolkit=10.2

conda install --yes -c conda-forge cmake=3.18 gtest==1.10.0=h0efe328_4 gmock cppzmq cython=0.29 openjdk=8.0 maven jpype1 netifaces pyhive tqdm ipywidgets

./build.sh -t disable-google-gs

wmalpica commented 3 years ago

resolved