amazon-science/ammo / ai_cli_session/.codex/skills/ammo
ammo
UnverifiedProfile and optimize GPU kernels for vLLM inference on NVIDIA GPUs. Use when targeting a specific model, hardware, dtype, parallelism, and workload to improve latency.
tests/test_mining_invalidated_schema.py
View on GitHub#!/usr/bin/env python3
# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
# SPDX-License-Identifier: CC-BY-NC-4.0
"""Z2-T1 / P4#7: schema legality of `mining_invalidated` (+ reason) round fields.
Pins the schema gap fix: the round object is `additionalProperties:false`, so
before P4#7 the engine reads `mining_invalidated` (ammo_state.advance, :1219)
that no schema-legal write could ever persist. This file adds the two optional
round properties and proves:
1. a round carrying BOTH fields (boolean `true` + reason string) validates and
round-trips through `advance --outcome EXHAUSTED` to 2_bottleneck_mining;
2. the load-bearing boolean requirement holds at BOTH layers — the schema
REJECTS the string form `"true"` (the `"type":"boolean"` declaration makes
the misuse impossible to persist), and the in-process `advance` on a dict
carrying the string routes to 3_debate (because `is True` is False for a
string).
ADDITIVE: this is a NEW file. It does NOT import, edit, mutate, or weaken the
frozen tests/test_ammo_state.py (TestAdvance, test_exhausted_mining_invalidated_at_mining,
test_exhausted_without_invalidation_goes_to_debate). Helpers are mirrored locally
on purpose so this file is self-contained.
Run: python3 -m pytest tests/test_mining_invalidated_schema.py -q
"""
from __future__ import annotations
import copy
import json
import subprocess
import sys
from pathlib import Path
_SCRIPTS_DIR = Path(__file__).resolve().parent.parent / "scripts"
_ENGINE = _SCRIPTS_DIR / "ammo_state.py"
_SCHEMA = Path(__file__).resolve().parents[3] / "schemas" / "state.schema.json"
sys.path.insert(0, str(_SCRIPTS_DIR))
import ammo_state # noqa: E402
# ─────────────────────────────────────────────
# Local fixtures (mirrored — NO import from the frozen test_ammo_state.py)
# ─────────────────────────────────────────────
def _round_skel(round_id, max_rounds=4):
return {
"round_id": round_id, "status": "IN_PROGRESS", "team_name": None,
"profiling_baseline_path": None,
"baseline": {"started_at": None, "completed_at": None, "e2e_latency": None, "per_bs_verdict": None},
"bottleneck_mining": {"started_at": None, "completed_at": None, "top_bottleneck_share_pct": None},
"debate": {"started_at": None, "completed_at": None, "candidates": [],
"rounds_completed": 0, "max_rounds": max_rounds, "selected_winners": []},
"parallel_tracks": {"started_at": None, "completed_at": None, "tracks": {}},
"integration": {"started_at": None, "completed_at": None, "status": "pending",
"passing_candidates": [], "failed_candidates": [], "selected_candidates": [],
"conflict_analysis": None, "combined_patch_branch": None, "combined_e2e_result": None,
"e2e_latency_combined": None, "per_bs_verdict": None, "commit_sha": None,
"final_decision": None, "resolver_invoked": None, "resolver_outcome": None,
"conflicting_tracks": None},
"campaign_eval": {"started_at": None, "completed_at": None},
"audit": {},
"shipped": [], "dropped": [],
"cumulative_speedup_after": None, "combined_e2e_speedup_x": None,
"combined_e2e_delta_pp": None, "note": None, "round_summary": None,
}
def _base_state():
"""Canonical valid v4.1 state — mirrors the bash harness base_v2 shape."""
return {
"target": {
"model_id": "test-model", "hardware": "H100", "dtype": "bf16",
"tp": 1, "dp": 1, "ep": 1, "component": "auto",
},
"session_id": None,
"gpu_resources": {
"gpu_count": 1, "gpu_model": "NVIDIA H100",
"memory_total_gib": 80.0, "cuda_visible_devices": "0",
},
"campaign": {
"schema_version": "4.1",
"status": "active",
"current_round": 1,
"current_stage": "1_baseline",
"config": {
"min_e2e_improvement_pct": 1.0,
"noise_tolerance_pct": 0.5,
"catastrophic_regression_pct": 5.0,
},
"cumulative_speedup_vs_round1": 1.0,
"round_1_baseline_latency_s": None,
"shipped_optimizations": [],
"agent_costs": [],
"rounds": [_round_skel(1)],
},
}
def _audited_round1_state():
"""Round 1 at 7_campaign_eval with a full audit chain (incl stage_67) so the
new-round-start gate is satisfied after advance creates round 2. Mirrored
locally; NOT imported from the frozen test."""
st = _base_state()
st["campaign"]["current_stage"] = "7_campaign_eval"
st["campaign"]["rounds"][0]["audit"] = {
"stage_1": {"passed_at": "2026-06-08T14:56:31Z"},
"stage_2": {"passed_at": "2026-06-08T15:48:49Z"},
"stage_45": {"passed_at": "2026-06-08T18:26:35Z"},
"stage_67": {"passed_at": "2026-06-08T18:42:23Z"},
}
return st
def _write_state(tmp_path, state):
"""Write a state.json inside a .codex/schemas-walkable artifact dir."""
root = tmp_path
(root / ".git").mkdir(exist_ok=True)
(root / ".codex" / "schemas").mkdir(parents=True, exist_ok=True)
(root / ".codex" / "schemas" / "state.schema.json").write_text(
_SCHEMA.read_text(encoding="utf-8"), encoding="utf-8")
art = root / "kernel_opt_artifacts" / "t"
art.mkdir(parents=True, exist_ok=True)
sp = art / "state.json"
sp.write_text(json.dumps(state, indent=2), encoding="utf-8")
return sp
def _run(*argv):
return subprocess.run([sys.executable, str(_ENGINE), *argv],
capture_output=True, text=True)
def _schema_validates(state):
"""In-process schema check using the same Draft202012Validator the engine uses."""
schema = json.loads(_SCHEMA.read_text(encoding="utf-8"))
errs = ammo_state.schema_errors(state, schema)
# schema_errors returns None only when jsonschema is unavailable (not here)
assert errs is not None, "jsonschema must be present to run this test"
return errs == []
# ─────────────────────────────────────────────
# Tests
# ─────────────────────────────────────────────
def test_round_with_both_fields_validates(tmp_path):
"""A round carrying mining_invalidated=true + reason string is schema-legal
and passes the engine `validate` verb (exit 0 / PASS)."""
st = _base_state()
st["campaign"]["rounds"][0]["mining_invalidated"] = True
st["campaign"]["rounds"][0]["mining_invalidated_reason"] = "wrong component attribution"
sp = _write_state(tmp_path, st)
r = _run("validate", "--state", str(sp))
assert r.returncode == 0, r.stdout + r.stderr
assert "PASS" in r.stdout, r.stdout
def test_advance_exhausted_mining_invalidated_file_roundtrip(tmp_path):
"""advance --outcome EXHAUSTED on an audited round-1 state carrying
mining_invalidated=true persists (validates before write) and routes the new
round to 2_bottleneck_mining (re-mine). The written doc still validates."""
st = _audited_round1_state()
st["campaign"]["rounds"][0]["mining_invalidated"] = True
st["campaign"]["rounds"][0]["mining_invalidated_reason"] = "diagnosis was wrong"
sp = _write_state(tmp_path, st)
r = _run("advance", "--state", str(sp), "--outcome", "EXHAUSTED")
assert r.returncode == 0, r.stdout + r.stderr
doc = json.loads(sp.read_text())
assert doc["campaign"]["current_stage"] == "2_bottleneck_mining", doc["campaign"]["current_stage"]
assert doc["campaign"]["rounds"][0]["status"] == "EXHAUSTED"
# the written doc (now carrying the optional fields) re-validates clean
assert _schema_validates(doc), "advanced doc with mining_invalidated must validate"
def test_string_true_is_rejected_and_does_not_trigger_remine(tmp_path):
"""Boolean-sensitivity guard (team-lead ruling 3), two assertions:
(i) SCHEMA rejects the string form: mining_invalidated="true" (string) fails
validate — the "type":"boolean" declaration makes the misuse impossible
to persist (the strongest guarantee).
(ii) ENGINE `is True` branch: advance(<dict-with-string-true>, EXHAUSTED)
routes to 3_debate, NOT 2_bottleneck_mining, because
`cur.get("mining_invalidated") is True` (ammo_state.py :1219) is False
for a string.
"""
# (i) schema rejects the string form
st = _base_state()
st["campaign"]["rounds"][0]["mining_invalidated"] = "true" # string, not bool
sp = _write_state(tmp_path, st)
r = _run("validate", "--state", str(sp))
assert r.returncode == 1, r.stdout + r.stderr
assert "PASS" not in r.stdout, r.stdout
assert not _schema_validates(st), "string 'true' must FAIL boolean type validation"
# (ii) in-process advance with the string routes to 3_debate (is True is False)
st2 = _base_state()
st2["campaign"]["current_stage"] = "7_campaign_eval"
st2["campaign"]["rounds"][0]["mining_invalidated"] = "true" # string
transitions = ammo_state.load_transitions()
_, new_stage = ammo_state.advance(copy.deepcopy(st2), "EXHAUSTED", transitions)
assert new_stage == "3_debate", new_stage
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |