Repository navigation
WIP: SEP 52 milestone 1 — native _testing module; remove Batch and every group path - #853
Merged
Merged
Conversation
…e 1) Delete the fixed-schedule Batch scan executor and the group machinery behind it on both backends and in the driver: driver/executors/batch.py, driver/batch.py (compatible_groups), driver/_scan_batch.py and the run_batch hook; Backend.max_batch, Backend.prepare_group and the BatchBackend protocol; GroupSession, PreparedGroup, check_group, check_member_ticket, start_group, GroupBatch and ExportTicket.member; select_batch; the reference's HostGroup, HostGroupSession and NumpyBatchBackend; solprop_native.group and its counters; and the group_batch benchmark case. StatsSpec.same_tables, whose only caller was the group's compatibility check, goes too. The Backend protocol keeps SEP 52 section 5.1's members. The leading unit axis of the trajectory's arrays stays for a later phase. Group-only tests are deleted; tests that used a group to reach a member's behaviour run singly. The scan build helper moves to tests/builders.py. The user guide, the API page, the developer pages and the structure figures drop Batch, pending changelog fragments that announced Batch features are corrected, and a Removed fragment names Local, Pool, Queue and Slurm as replacements. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…s to _testing (SEP 52 phase 0) Delete the bindings nothing reaches (table_evaluate, table_rate, set_flow_actions, shutdown_runtime) and the resident Adams engine's per-step adapter, which the compiled march (SEP 39) replaced: its trials, accepts, start-up, holds and dense observations, with the bindings only they reached (adams_fused_trial_at, adams_fused_accept, adams_observe, adams_exponential_observe). _ResidentAdams keeps its construction, resume and snapshot sizing; the timing probe's paths stay until phase 4. Move every entry point only solprop's tests reach into a private `_testing` submodule built into the extension: the scripted marches and march runs, fault injection, the monitor and transverse-edge owners and watch functions, the engine-level parity reads, the state-level statistics, the vector-maths and width helpers, and the runtime probes. Moved State methods take the state first. Repoint the tests; the resident-history helpers they used move to tests/sessions/runs.py over `_testing`. The `_testing` stub is hand-written (_core_testing.pyi); problem_stub.py --check holds its names to the module's. The wheel smoke test runs one of its checks. Release builds pin nanobind 3.1.0. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts: # tests/sessions/test_native_session_axis_plasma.py
… [skip ci] Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Documentation build overview
91 files changed ·
|
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
jtravs
marked this pull request as ready for review
November 16, 2026 03:58
Codex Review Summary: ✅ CompletedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Milestone 1 of SEP 52 (accepted in #850), tracked in #849. This is the first of four stacked pull requests (decision D7): later milestones are based on this branch. The details are in the SEP's Part II report for milestone 1.
Phase 0 — native deletions and the test module
_ResidentAdamsadapter, together with the bindings that only it reached. The timing probe's paths stay until phase 4.solprop_native._core._testing, a private submodule built into the same extension. It holds the 54 test-only entry points (traces, fault injection, owners, engine parity, statistics, helpers, runtime probes), and every test that calls them is repointed.Stategoes from 96 bound names to 55._testing. nanobind is pinned to 3.1.0.Phase 1 — remove
Batchand every group pathBatchexecutor,compatible_groups, the group service loop,GroupSession/PreparedGroup,HostGroup,NumpyBatchBackend,NativeGroupSession,select_batch,BatchBackend,Backend.max_batch/prepare_groupand thegroup_batchbenchmark.Removedchangelog fragment namesLocal,Pool,QueueandSlurm.Batchgroup #171 and SEP 50 follow-up: scans of multi-pass cells cannot use theBatchexecutor #828 as moot (to be closed at merge).Size: −9.7k / +0.6k lines over 137 files for phase 1. Python total 271,036 → 266,928.
Overlap with SEP 51 (#851): this branch touches
driver/_scan_run.pyand sometests/driverfiles that SEP 51 may also edit. The hunks don't overlap; no save-plane behaviour changes.Checks run locally: targeted tests on an M5 Max (about 3,000 tests across the touched modules, plus native tests); ruff, basedpyright, docstrings, codespell,
poe docs,poe size-check. One native test,test_band_plasma.py::…[19], fails locally on a thread-sizing assertion unrelated to this change; CI will confirm. The full gate, coverage and wheel lanes run here.🤖 Generated with Claude Code