[codex] Tie #1415 to IC6 tag-cooccurrence coverage - #1427

Refs #1415.

Summary

• Tie the existing IC6 tag-cooccurrence join+aggregation regression coverage to post-rerun residual issue [POST-RERUN] #880 residual: tag-cooccurrence join-aggregation bucket #1415 while preserving the original [FEA] GFQL residual row-bindings: tag cooccurrence join aggregation #1396 lineage.
• Update the unreleased changelog entry to cite [POST-RERUN] #880 residual: tag-cooccurrence join-aggregation bucket #1415 for the tag_cooccurrence_join_aggregation [FEA] GFQL bindings_table (row bindings) from chain using Yannakakis prune #880 residual lane.

Current Status

• Ready for normal review/approval; this PR is not a [POST-RERUN] #880 residual: tag-cooccurrence join-aggregation bucket #1415 closeout.
• [POST-RERUN] #880 residual: joined row projection bucket #1411 is closed by Add GFQL row join apply (#1411) #1426, and [POST-RERUN] #880 residual: joined row aggregation bucket #1413 is closed by Fix #1413 IC3 carried-row reentry #1428.
• This branch is rebased on origin/master at 2f83b27da8108aab6b2d4770bf31ac0d562a4cc0 with head 4dc17c6081089ba71d36d9507abda5f9f93a48cb.
• Latest check after fetching origin/master: branch is 1 ahead / 0 behind.
• PR CI is green after the force-push/rebase.
• Review loop converged with no findings.
• DGX RAPIDS 25.02 + 26.02 targeted cuDF validation passed on dgx-spark.

#1415 Closure

• Keep this as Refs #1415; do not close [POST-RERUN] #880 residual: tag-cooccurrence join-aggregation bucket #1415 from this PR.
• The remaining [POST-RERUN] #880 residual: tag-cooccurrence join-aggregation bucket #1415 acceptance is benchmark-side: fresh SNB IC6 evidence must show the lane no longer reports backend=gfql, status=partial, issue_ref=graphistry/pygraphistry#880.
• That work is tracked in graphistry/pyg-bench#12.
• The latest DGX SNB closeout attempt staged SF1 successfully, but did not produce a closeout artifact: the wrapper path hit nested SSH auth, and the direct RAPIDS 25.02 container path exited 137 before writing result.json / probe-results.json.
• Inspecting pyg-bench showed IC6 tag-cooccurrence GFQL is still classified as adapter_workaround, success_status: partial, and success_issue_refs: [graphistry/pygraphistry#880] by construction.

Validation

• ./bin/lint.sh CHANGELOG.md graphistry/tests/compute/gfql/cypher/test_lowering.py
• python3 -m pip install --target /tmp/pyg1415-deps 'lark>=1.1,<2'
• PYTHONPATH=/tmp/pyg1415-deps python3 -m pytest -q graphistry/tests/compute/gfql/cypher/test_lowering.py::test_issue_1396_issue_1415_tag_cooccurrence_join_aggregation_counts graphistry/tests/compute/gfql/cypher/test_lowering.py::test_issue_1000_ic6_compiles_after_phase6_connected_star_fanout graphistry/tests/compute/gfql/cypher/test_lowering.py::test_string_cypher_executes_issue_1000_ic6_exact_runtime_minimal graphistry/tests/compute/gfql/cypher/test_lowering.py::test_string_cypher_executes_scalar_prefix_reentry_connected_star_comma_fanout_grouped_count -> 4 passed
• PYTHONPATH=/tmp/pyg1415-deps python3 -m pytest -q graphistry/tests/benchmarks/gfql/test_benchmark_residual_triage.py::test_bucket_classification -> 1 passed
• git diff --check
• ssh dgx-spark 'cd /tmp/pygraphistry-issue-1415 && MATRIX_CELLS="25.02:gfql 26.02:gfql" WITH_IMAGE_BUILD=0 WITH_GPU=1 TEST_FILES="graphistry/tests/compute/gfql/cypher/test_lowering.py::test_issue_1396_issue_1415_tag_cooccurrence_join_aggregation_counts_on_cudf" ./docker/test-rapids-official-matrix.sh' -> 25.02/gfql PASS, 26.02/gfql PASS
• gh pr checks 1427 -> green after latest force-push/rebase

Local Caveats

• System python3 was missing lark; direct install is blocked by PEP 668, so validation used a /tmp target install.
• Local cuDF tests cannot run because this host has cuDF importable but no CUDA-capable device (cudaErrorNoDevice).