Adopt cuco's parametric bloom filter policy in Parquet and streaming

[PointKernel]

Description

This PR migrates the Parquet reader and the streaming bloom filter onto cuco's new parametric_filter_policy (bumped via rapidsai/rapids-cmake#1044). The Parquet reader defines a local arrow_filter_policy alias hashing keys with XXHash_64, and streaming uses the policy directly with identity_hash; no shared cudf policy header is added. Depends on NVIDIA/cuCollections#825, which drops the hasher argument_type requirement so XXHash_64 can be used without an adapter.

Checklist

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

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[PointKernel]

/ok to test 0c16a97

[PointKernel]

/ok to test 81828e6

[PointKernel]

/ok to test ba07b5c

[PointKernel]

/ok to test 189ac84

[PointKernel]

/ok to test 763fb86

[PointKernel]

This is ready for review and the CMake changes will be reverted once the upstream #1044 gets in.

[mhaseeb123]

Awesome work. LGTM (with test/debug artifacts removed of course)

[wence-]

With the caveat around the streaming detail usage.

[coderabbitai]

📝 Walkthrough

Summary by CodeRabbit

• New Features

  • Bloom filter handling now supports a refreshed configuration for Parquet and join-related filtering.

• Bug Fixes

  • Improved consistency in bloom filter behavior across reading, testing, and join workflows.
  • Updated copyright notices and test setup to align with the latest project configuration.

Walkthrough

Replaces cudf's custom arrow_filter_policy header with cuco::parametric_filter_policy instantiations across the Parquet bloom filter reader, streaming device bloom filter, mark join, and tests. Also removes arrow_filter_policy.cuh entirely and temporarily pins rapids-cmake repo/branch to a test fork for validation.

Changes

Bloom filter policy migration

Layer / File(s)	Summary
Arrow filter policy migration in Parquet reader cpp/src/io/parquet/arrow_filter_policy.cuh, cpp/src/io/parquet/bloom_filter_reader.cu	Deletes the custom arrow_filter_policy header and introduces an in-file alias in bloom_filter_reader.cu using cuco::parametric_filter_policy with XXHash_64, matching Arrow's Block-Split Bloom Filter layout.
Streaming device bloom filter policy update cpp/libcudf_streaming/src/detail/device_bloom_filter.cu	Updates BloomFilterRefType to use cuco::parametric_filter_policy with cuco::identity_hash<KeyType> instead of the removed arrow_filter_policy.
Mark join bloom filter policy simplification cpp/src/join/mark_join.cuh	Simplifies bloom_filter_policy_type to cuco::default_filter_policy<hash_value_type>, removing the explicit identity-hash instantiation.
Parquet bloom filter test policy update cpp/tests/io/parquet_bloom_filter_test.cu	Replaces policy_type with an explicit cuco::parametric_filter_policy<XXHash_64<key_type>, ...> instantiation and updates includes accordingly.
Rapids-cmake test pin cmake/rapids_config.cmake	Overrides rapids-cmake-repo and rapids-cmake-branch to point to a test fork/branch (cuco-new-filter) for validation.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Possibly related PRs

• rapidsai/cudf#22415: Introduced the arrow_filter_policy.cuh header that this PR removes and replaces with cuco::parametric_filter_policy in the same reader and tests.

Suggested labels: cuIO, cuco, feature request

Suggested reviewers: mythrocks, bdice, qbacpey, Matt711, vyasr

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main change: adopting cuco's parametric bloom filter policy in Parquet and streaming.
Description check	✅ Passed	The description matches the changeset and explains the Parquet and streaming bloom filter policy migration.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@cmake/rapids_config.cmake`:
- Around line 37-40: Remove the temporary rapids-cmake test pin so main does not
always fetch from the personal fork/branch. The unconditional assignments to
rapids-cmake-repo and rapids-cmake-branch currently bypass the existing NOT
rapids-cmake-branch guard and override caller settings; delete or revert this
test block before merge. Use the rapids-cmake configuration logic in
rapids_config.cmake as the place to restore the normal override behavior.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 6ff23f16-b06c-4df9-9ea0-01cde3c29991

📥 Commits

Reviewing files that changed from the base of the PR and between 0732eaa and f7c48a6.

📒 Files selected for processing (6)

• cmake/rapids_config.cmake
• cpp/libcudf_streaming/src/detail/device_bloom_filter.cu
• cpp/src/io/parquet/arrow_filter_policy.cuh
• cpp/src/io/parquet/bloom_filter_reader.cu
• cpp/src/join/mark_join.cuh
• cpp/tests/io/parquet_bloom_filter_test.cu

💤 Files with no reviewable changes (1)

• cpp/src/io/parquet/arrow_filter_policy.cuh

[PointKernel]

/merge