Install linux with mlx[cuda] and mlx[cpu] (#2356)

* install linux with mlx[cuda] and mlx[cpu]

* temp for testing

* cleanup circle, fix cuda repair

* update circle

* update circle

* decouple python bindings from core libraries

python/scripts/repair_cuda.sh

@@ -1,17 +1,23 @@
 #!/bin/bash
 
 auditwheel repair dist/* \
-  --plat manylinux_2_35_x86_64 \
+  --plat manylinux_2_39_x86_64 \
   --exclude libcublas* \
-  --exclude libnvrtc*
+  --exclude libnvrtc* \
+  -w wheel_tmp
 
-cd wheelhouse
+
+mkdir wheelhouse
+cd wheel_tmp
 repaired_wheel=$(find . -name "*.whl" -print -quit)
 unzip -q "${repaired_wheel}"
-core_so=$(find mlx -name "core*.so" -print -quit)
-rpath=$(patchelf --print-rpath "${core_so}")
-rpath=$rpath:\$ORIGIN/../nvidia/cublas/lib:\$ORIGIN/../nvidia/cuda_nvrtc/lib
-patchelf --force-rpath --set-rpath "$rpath" "$core_so"
+rm "${repaired_wheel}"
+mlx_so="mlx/lib/libmlx.so"
+rpath=$(patchelf --print-rpath "${mlx_so}")
+base="\$ORIGIN/../../nvidia"
+rpath=$rpath:${base}/cublas/lib:${base}/cuda_nvrtc/lib
+patchelf --force-rpath --set-rpath "$rpath" "$mlx_so"
+python ../python/scripts/repair_record.py ${mlx_so}
 
 # Re-zip the repaired wheel
-zip -r -q "${repaired_wheel}" .
+zip -r -q "../wheelhouse/${repaired_wheel}" .