2c3c93d940
Give the producer lane (mode 2, surface-gated) its designed shape + the missing ② Serve link (cockpit realignment §3.3 / §9.4 R3). - ModeSwitcher "Validate" → "Bring & Validate" (mode index 2 unchanged). Mode 2 is now an ordered 5-stage pipeline: ① Bring (LaneBringPane, reuses byo_check) · ② Serve · ③ Gate (ValidateRunPane) · ④ Measure (ValidateEvidencePane) · ⑤ Promote (LanePromotePane). - ② Serve (new): generate_compose() shells generate-compose.sh --profile <slug> --out <tmp> + reads back the YAML; serve_generated() builds a reconcile-gated serve ActionPlan; action_serve_untested previews the compose VERBATIM (UntestedComposePreviewScreen) and serves it through the SAME ConfirmActionScreen → reconcile gate (never auto-fired). - Relocated [P] Promote (Run→lane) and [v] c3t Evaluate (Operate→lane); both now mode-2 + in _PRODUCER_ONLY. HelpScreen surface-threaded (consumer help omits the lane / Promote / Evaluate). Verified by an implement→gate→repair→verify→critic workflow; folded the critic's must-fix: - HIGH: [v] Evaluate had no data in the lane (_target_obj was primed only by the Operate poll). load_estate() now also runs on lane entry + lane refresh, so the lane-native path (key 3, never visiting Operate) has the live target. - MED: ② Serve was serving the catalog/sibling profile's weights under a brought- model "untested" badge (generate-compose.sh has no weights-swap). Made HONEST: badge/modal/docstrings state it serves a verbatim REPRODUCTION of the resolved catalog profile, and it no longer silently falls back to a generic profile (notifies if none resolves). Serving the BROUGHT model (pull-to-disk + a generate-compose.sh --repo extension) is a tracked follow-up. - MED: generate_compose's temp compose (c3-genc-*.yml at the repo root, so the ../ mounts resolve) is now gitignored — no longer exposed to git add -A. Deferred to R4 (the fold subagent hit a session limit before these LOW items): the temp unlink-on-decline (disk cleanup; git-pollution already fixed via gitignore), reverting the filter-focus footer gate, and two test-assert tightenings. Suite green: 342 passed. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
68 lines
2.1 KiB
Plaintext
68 lines
2.1 KiB
Plaintext
# Model weights — never commit. Default download dir is models-cache/ (set
|
|
# via MODEL_DIR env var if you want a different path).
|
|
models-cache/
|
|
*.safetensors
|
|
*.gguf
|
|
*.bin
|
|
|
|
# Genesis patches fetched by setup.sh (upstream, not ours to vendor).
|
|
# Lives at models/<model>/vllm/patches/genesis/ in the new layout.
|
|
# No trailing slash — also matches local symlinks pointing at out-of-tree clones.
|
|
models/*/vllm/patches/genesis
|
|
|
|
# v0.8.0 [E] Pull-Emit-Derived runtime capture artifacts (the §6 capture
|
|
# points + manifest; consumed by the later [F] Loop phase). Per-run output,
|
|
# never committed.
|
|
.pull-captures/
|
|
|
|
# serve-cockpit producer lane ② Serve — temp composes generated by
|
|
# generate-compose.sh (CockpitData.generate_compose). Written under the repo
|
|
# root so their relative ../ mounts resolve; cleaned up on serve/decline, but
|
|
# gitignored so a crash-leftover never lands in `git add -A`.
|
|
c3-genc-*.yml
|
|
|
|
# Python
|
|
__pycache__/
|
|
*.pyc
|
|
*.pyo
|
|
*.egg-info/
|
|
.venv/
|
|
venv/
|
|
|
|
# Editor
|
|
.vscode/
|
|
.idea/
|
|
*.swp
|
|
.DS_Store
|
|
|
|
# Docker compose state
|
|
**/compose/*.log
|
|
|
|
# Local env overrides
|
|
.env
|
|
*.env.local
|
|
|
|
# Local safety backups of compose files (e.g. docker-compose.yml.default)
|
|
**/compose/*.default
|
|
|
|
# Patched vLLM source clone (volume-mounted by dual-card composes)
|
|
vllm-src/
|
|
|
|
# Internal diagnostics / feasibility memos — kept local, not published.
|
|
# Per-model published diagnostics live at models/<name>/vllm/diagnostics/.
|
|
docs/diagnostics/
|
|
|
|
# Bench output. Most runs are intermediate investigations (cliff probes, soak
|
|
# validation matrices, residency instrumentation) — those live in issue threads
|
|
# and gitignored docs/diagnostics/. Bench evidence backing numbers in BENCHMARKS.md
|
|
# / STRUCTURED_COT.md / CHANGELOG is tracked via the negations below; future
|
|
# grammar-ab-* and grammar-full-* runs will auto-track when committed.
|
|
results/*
|
|
!results/grammar-ab-*/
|
|
!results/grammar-full-*/
|
|
!results/lucebox-*/
|
|
!results/v0.20-migration/
|
|
# #252: curated quality-baseline corpus (committed; the trusted n>=3 aggregates
|
|
# `quality-baseline.sh` diffs fresh runs against). Run output elsewhere stays ignored.
|
|
!results/baselines/
|