Add rapids-pip-retry across RAPIDS

[gforsyth]

We've observed several failures across CI, usually related to large wheels, where a network hiccup interrupts the download but pip attempts to continue with installation with the partially downloaded file, causing a hash mismatch error.

It looks something like this in CI:

   Collecting nvidia-cublas-cu12 (from libraft-cu12==25.2.*,>=0.0.0a0)
    Downloading https://pypi.nvidia.com/nvidia-cublas-cu12/nvidia_cublas_cu12-12.8.3.14-py3-none-manylinux_2_27_aarch64.whl (604.9 MB)
       ━━━━━━━━━━━━━━━━━━━━━                 350.2/604.9 MB 229.2 MB/s eta 0:00:02
  ERROR: THESE PACKAGES DO NOT MATCH THE HASHES FROM THE REQUIREMENTS FILE. If you have updated the package versions, please update the hashes. Otherwise, examine the package contents carefully; someone may have tampered with them.
      nvidia-cublas-cu12 from https://pypi.nvidia.com/nvidia-cublas-cu12/nvidia_cublas_cu12-12.8.3.14-py3-none-manylinux_2_27_aarch64.whl#sha256=93a4e0e386cc7f6e56c822531396de8170ed17068a1e18f987574895044cd8c3 (from libraft-cu12==25.2.*,>=0.0.0a0):
          Expected sha256 93a4e0e386cc7f6e56c822531396de8170ed17068a1e18f987574895044cd8c3
               Got        849c88d155cb4b4a3fdfebff9270fb367c58370b4243a2bdbcb1b9e7e940b7be

This error doesn't get flagged by pip as "retryable", so settings like PIP_RETRIES doesn't impact it.

Since these failures are intermittent, we're attempting to mitigate them by wrapping pip with a utility that retries the failing command if it detects one of these hash errors (see rapidsai/gha-tools#132).

RAPIDS repos to add rapids-pip-retry to:

Tasks

Preview Give feedback

1. cucim (Use rapids-pip-retry in CI jobs that might need retries cucim#824)
2. cudf (Use rapids-pip-retry in CI jobs that might need retries cudf#17920)
3. cugraph (Use rapids-pip-retry in CI jobs that might need retries cugraph#4913)
4. cugraph-gnn (Use rapids-pip-retry in CI jobs that might need retries cugraph-gnn#133)
5. cuml (Use rapids-pip-retry in CI jobs that might need retries cuml#6293)
6. cuspatial (Use rapids-pip-retry in CI jobs that might need retries cuspatial#1533)
7. cuvs (Use rapids-pip-retry in CI jobs that might need retries cuvs#648)
8. cuxfilter (Use rapids-pip-retry in CI jobs that might need retries cuxfilter#655)
9. dask-cuda (Use rapids-pip-retry in CI jobs that might need retries dask-cuda#1443)
10. kvikio (Use rapids-pip-retry in CI jobs that might need retries kvikio#619)
11. nx-cugraph (Use rapids-pip-retry in CI jobs that might need retries nx-cugraph#80)
12. pynvjitlink (Use rapids-pip-retry in CI jobs that might need retries pynvjitlink#125)
13. raft (Use rapids-pip-retry in CI jobs that might need retries raft#2571)
14. rmm (Use rapids-pip-retry in CI jobs that might need retries rmm#1814)
15. ucx-py (Use rapids-pip-retry in CI jobs that might need retries ucx-py#1114)
16. ucxx (Use rapids-pip-retry in CI jobs that might need retries ucxx#366)