Skip to content

[WIP] Refs #1416 recursive ancestor residual closeout - #1430

Closed
lmeyerov wants to merge 1 commit into
masterfrom
codex/issue-1416-recursive-ancestor-residual-closeout
Closed

[WIP] Refs #1416 recursive ancestor residual closeout#1430
lmeyerov wants to merge 1 commit into
masterfrom
codex/issue-1416-recursive-ancestor-residual-closeout

Conversation

@lmeyerov

@lmeyerov lmeyerov commented May 14, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Refs #1416.

Scope

  • Verification/evidence PR for the recursive-ancestor residual closeout.
  • No pygraphistry source change: this branch is an empty anchor commit on latest master.
  • Heavy validation ran on dgx-spark; no heavy local benchmark/GPU work.

Current Evidence

  • [codex] Add #1416 recursive ancestor row-join regression #1423 added pygraphistry recursive ancestor row-join regression coverage.
  • graphistry/pyg-bench#10 merged and removed the SNB IS6 / message-forum adapter workaround, replacing it with native rows() recursive ancestor traversal.
  • Fresh DGX staged snapshot for this PR:
  • Focused pyg-bench unit validation on DGX:
    • PYTHONPATH=. /home/lmeyerov/.local/bin/uv run --with pytest pytest -q tests/test_snb_interactive_runner.py -k "message_forum_native_rows or recent_replies_native_rows"
    • 2 passed, 27 deselected in 0.94s
  • Fresh DGX SF0.1 dataset-backed direct probe:
    • artifact: /tmp/pyg-bench-1416-closeout-main/results/runs/dgx-direct-issue-1416-message-forum-sf01-2026-05-14-pr1430
    • query_id=message-forum, upstream_query=interactive-short-6, backend=gfql
    • status=ok, semantic_scope=native_rows_recursive_ancestor_join
    • issue_refs=[], rows_returned=1, probes_partial=0

Remaining Gate

  • Fresh SF1/latest direct probe was attempted on DGX:
    • output dir: results/runs/dgx-direct-issue-1416-message-forum-sf1-2026-05-14-pr1430
    • command exited 137 before writing artifacts.
    • start memory: ~31GiB available, swap 9.8GiB/15GiB used.
    • post-failure memory: ~33GiB available, swap 12GiB/15GiB used.
  • [POST-RERUN] #880 residual: recursive-ancestor row-join bucket #1416 should remain open unless the coordinator accepts the fresh SF0.1 direct probe as the lower-memory equivalent requested by the issue.

Validation

  • DGX focused pyg-bench unit validation.
  • DGX fresh lower-memory direct dataset probe is clean.
  • DGX SF1 attempted; blocked by exit 137 before artifacts.
  • Local focused residual-triage test: python3 -m pytest -q graphistry/tests/benchmarks/gfql/test_benchmark_residual_triage.py -> 5 passed.
  • Review skill converged: plans/issue-1416/final-report.md.
  • CI green.
  • Latest origin/master drift checked: branch is ahead 1, behind 0.

@lmeyerov lmeyerov mentioned this pull request May 14, 2026
Closed
@lmeyerov

lmeyerov commented May 15, 2026

Copy link
Copy Markdown
Contributor Author

Coordinator decision recorded on #1416: the fresh DGX SF0.1 direct probe is accepted as the lower-memory equivalent for #1416.

This PR should not stay parked. It has no code diff and CI is green. Next action is one of:

  1. Close this PR unmerged as a receipt-only anchor and close [POST-RERUN] #880 residual: recursive-ancestor row-join bucket #1416 from the issue receipt, or
  2. Undraft, get normal approval, merge as an audit artifact, then close [POST-RERUN] #880 residual: recursive-ancestor row-join bucket #1416.

Preferred path: close unmerged unless the team specifically wants the PR audit artifact. No more worker time should be spent waiting for SF1 unless there is a concrete DGX memory-window plan.

@lmeyerov

lmeyerov commented May 15, 2026

Copy link
Copy Markdown
Contributor Author

Closing this PR unmerged as the receipt-only closeout path for #1416.

Coordinator direction allowed either:

Chosen path: close unmerged. This PR has no source diff, only an empty anchor commit, and the evidence is already preserved in the PR body plus #1416 comments. Merging would add an empty audit commit to master and now require a rebase after newer master movement, without changing pygraphistry behavior.

Closeout evidence:

  • graphistry/pyg-bench#10 merged and removed the IS6 / message-forum adapter workaround.
  • Fresh DGX SF0.1 dataset-backed direct probe for interactive-short-6 / message-forum passed with backend gfql, status=ok, semantic_scope=native_rows_recursive_ancestor_join, issue_refs=[], rows_returned=1, probes_partial=0.
  • Fresh SF1 direct probe was attempted and exited 137 before artifacts under DGX memory/swap pressure; this is recorded for traceability.
  • PR CI was green.
  • Review skill converged with no findings.

Next action: close #1416 as completed per coordinator/owner closeout path.

@lmeyerov lmeyerov closed this May 15, 2026
This was referenced May 15, 2026
Open
Closed
Open
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@lmeyerov