Skip to content

[POST-RERUN] #880 residual: reply-author row-shaping bucket #1412

Description

@lmeyerov

Summary

Post-rerun residual split from #880: reply-author row-shaping join bucket was workaround-backed partial.

Final status

  • [GFQL] Lock reply-author row-shaping coverage #1425 merged 2026-05-10 PT: locked reply-author row-shaping coverage.
  • graphistry/pyg-bench#9 merged 2026-05-11 UTC at 518691b5220052cec0889a64deb3148598ee99ed: inspected/removed the remaining pyg-bench reply-author adapter workaround and added CHANGELOG.md.
  • DGX pinned-fixture sanity rerun passed after Report server-side ETL errors to client  #9 merged: results/runs/dgx-spark-issue-1412-testdata-absolute-closeout, run_id=snb-interactive-0dd69802e9, status=ok, probes_total=4, probes_passed=4, probes_partial=0.
  • DGX staged sf0.1 residual closeout passed: results/runs/dgx-spark-issue-1412-sf01-gfql-closeout, run_id=snb-interactive-c8dee7f018, status=ok, probes_total=2, probes_passed=2, probes_partial=0.
  • Target lanes no longer emit gfql partial with #880 in issue_refs.

Closeout receipt

  • message-replies / interactive-short-7 / GFQL: status=ok, issue_refs=[], semantic_scope=native_rows_reply_author_projection, rows_returned=13.
  • recent-replies / interactive-complex-8 / GFQL: status=ok, issue_refs=[], semantic_scope=official_query_text, rows_returned=20.
  • Bootstrap: overall_status=ok, upstream_status=ok, dataset_status=ok.

Earlier blocker resolved

  • SF1 remained too heavy for this closeout shape and returned 137 before artifacts.
  • The missing sf0.1 DGX dataset cache was staged from official LDBC downloads, enabling the final staged residual receipt above.

Repro Receipt

  • Source rerun set: results/runs/local-snb-interactive-closeout-2026-05-09
  • Residual report filter: backend=gfql, status=partial, issue_ref=graphistry/pygraphistry#880
  • Bucket: reply_author_row_shaping_join
  • Lane count: 2

Affected upstream lanes

  • interactive-complex-8 / recent-replies
  • interactive-short-7 / message-replies

Scope

Done when

Related

Activity

  1. lmeyerov commented on May 11, 2026

    @lmeyerov
    ContributorAuthor

    PR #1425 merged on 2026-05-11 UTC (2026-05-10 PT): #1425

    What landed:

    • native GFQL rows() regression coverage for the two affected reply-author row-shaping shapes:
      • IC8 / recent-replies
      • IS7 / message-replies
    • parity checks for bare native rows() and explicit rows(binding_ops=...), locking the pygraphistry-side behavior needed to retire adapter-local reply-author joins.

    Validation before merge:

    • targeted local pytest passed
    • ruff passed
    • PR CI passed, including Read the Docs

    Keeping this issue open for the remaining acceptance receipt in the issue body: pyg-bench still needs to remove the adapter workaround for these lanes and run a fresh residual report showing they no longer emit gfql partial with graphistry/pygraphistry#880.

  2. lmeyerov commented on May 11, 2026

    @lmeyerov
    ContributorAuthor

    #1412 direct closeout update

    Worked directly under #1412; no new issue filed.

    Implementation PR:

    What changed:

    • IC8 recent-replies: removed the #880-attributed adapter-local reply-author join path and moved GFQL to native rows() multi-alias materialization with projection, ordering, and limit.
    • IS7 message-replies: moved reply/comment/author projection to native rows(); only the optional KNOWS boolean remains as a staged person_knows postprocess because GFQL does not yet expose a first-class optional pattern-existence row expression.
    • Updated pyg-bench translation/readiness metadata so these lanes no longer read as #880 partial reply-author adapter-workaround lanes after PR merge.

    Validation:

    • uv run --extra dev python -m pytest -q tests/test_snb_interactive_runner.py -k 'recent_replies_native_rows or message_replies_gfql_native_rows' -> 2 passed
    • uv run --extra dev python -m pytest -q tests/test_snb_interactive_runner.py -> 28 passed
    • uv run --extra dev python -m pytest -q -> 103 passed
    • PYTHONPATH=/home/lmeyerov/Work/pygraphistry uv run --extra dev python -m pytest -q /home/lmeyerov/Work/pygraphistry/graphistry/tests/test_compute_chain.py -k issue_1412 -> 2 passed
    • YAML parse check for configs/analysis/snb-interactive-translation.yaml -> ok
    • git diff --check -> clean

    Rerun receipts:

    • IC8 command: uv run --extra dev python scripts/run_suite.py --suite snb-interactive --config configs/suites/snb-interactive-ic8-conformance-sf1.yaml --output-dir results/runs/local-issue-1412-ic8-closeout
    • IC8 run id: snb-interactive-0691677af8
    • IC8 result: adapter_missing; no probes executed because this fresh pyg-bench checkout is missing the pinned upstream Cypher query path and SF1 dataset.
    • IS7 command: uv run --extra dev python scripts/run_suite.py --suite snb-interactive --config configs/suites/snb-interactive-is7-conformance-sf1.yaml --output-dir results/runs/local-issue-1412-is7-closeout
    • IS7 run id: snb-interactive-9645ae5af8
    • IS7 result: adapter_missing; no probes executed because this fresh pyg-bench checkout is missing the pinned upstream Cypher query path and SF1 dataset.

    Remaining gate:

  3. lmeyerov commented on May 11, 2026

    @lmeyerov
    ContributorAuthor

    #1412 PR #9 CI follow-up

    Pushed a second commit to graphistry/pyg-bench#9:

    • 0f14382 — mechanical lint cleanup exposed by the first PR smoke run.

    Current PR status:

    The residual rerun gate remains unchanged: merge graphistry/pyg-bench#9, then rerun on an environment with the pinned upstream query checkout and SF1 dataset staged before closing #1412 / clearing #880.

  4. lmeyerov commented on May 11, 2026

    @lmeyerov
    ContributorAuthor

    Post-merge closeout update:

    • graphistry/pyg-bench#9 merged 2026-05-11 UTC at 518691b5220052cec0889a64deb3148598ee99ed.
    • That PR removed the remaining reply-author adapter workaround and added the pyg-bench CHANGELOG.md entry.
    • DGX pinned-fixture sanity rerun passed: results/runs/dgx-spark-issue-1412-testdata-absolute-closeout, run_id=snb-interactive-0dd69802e9, status=ok, probes_total=4, probes_passed=4, probes_partial=0.
    • Covered lanes/backends: gfql + cypher for message-replies and recent-replies.

    Keeping #1412 open for now. The full staged residual rerun is still not complete: SF1 IC8 died before probes with returncode=137, and the expected DGX sf0.1 dataset cache was missing for the closeout config.

  5. lmeyerov commented on May 11, 2026

    @lmeyerov
    ContributorAuthor

    Final #1412 closeout receipt:

    • Adapter workaround removal is done via graphistry/pyg-bench#9, merged at 518691b5220052cec0889a64deb3148598ee99ed.
    • Staged missing sf0.1 SNB cache on dgx-spark from official LDBC downloads.
    • DGX staged sf0.1 residual closeout passed: results/runs/dgx-spark-issue-1412-sf01-gfql-closeout, run_id=snb-interactive-c8dee7f018, status=ok, probes_total=2, probes_passed=2, probes_partial=0.
    • message-replies / interactive-short-7 / GFQL: status=ok, issue_refs=[], semantic_scope=native_rows_reply_author_projection, rows_returned=13.
    • recent-replies / interactive-complex-8 / GFQL: status=ok, issue_refs=[], semantic_scope=official_query_text, rows_returned=20.
    • Bootstrap was clean: overall_status=ok, upstream_status=ok, dataset_status=ok.

    Closing #1412 as completed.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions