Exclude nanoarrow and flatbuffers from installation

[vyasr]

Description

This change helps shrink RAPIDS wheels. It should not affect Spark builds since those use the build directory of cudf and statically link in those components to its final binary.

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[vyasr]

CC @jameslamb

[vyasr]

For the record, I spot-checked the wheels, one each for libcudf/pylibcudf/cudf:

(main) dt08% unzip -l cudf_cu11-24.12.0a344-cp310-cp310-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl | grep "flatbuffers"
(main) dt08% unzip -l cudf_cu11-24.12.0a344-cp310-cp310-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl | grep "nanoarrow"  
(main) dt08% unzip -l pylibcudf_cu12-24.12.0a344-cp311-cp311-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl| grep "flatbuffers"
(main) dt08% unzip -l pylibcudf_cu12-24.12.0a344-cp311-cp311-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl| grep "nanoarrow"  
(main) dt08% unzip -l libcudf_cu12-24.12.0a344-py3-none-manylinux_2_28_x86_64.whl | grep "flatbuffer"
(main) dt08% unzip -l libcudf_cu12-24.12.0a344-py3-none-manylinux_2_28_x86_64.whl | grep "nanoarrow" 

[vyasr]

/merge

[jameslamb]

Oh awesome, glad this was really all it took. Thank you!