Datasets:
File size: 4,487 Bytes
0e239c6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 | """Run a command contract inside its prepared evaluation container."""
import argparse
import json
import os
from pathlib import Path
import subprocess
import time
def execute(contract, output, environment_setup=None):
if contract["kind"] != "command_contract":
raise ValueError("Native contracts require a task-specific adapter")
output = Path(output)
output.mkdir(parents=True, exist_ok=True)
commands = contract["commands"]
def ordered_by(field):
return sorted(enumerate(commands), key=lambda item: (
item[1][field] if item[1].get(field) is not None else item[0], item[0]))
source_order = ordered_by("setup_order")
ordered = ordered_by("evaluation_setup_order")
records = []
infrastructure_failed = False
tests_failed = False
def command(entry, phase, text, cwd=None):
if not text:
return True
environment = dict(os.environ)
environment.update({key: str(value) for key, value in (entry.get("environment") or {}).items()})
text = text.replace("{{WORLD}}", contract["evaluation_world"])
working = cwd or entry.get("test_cwd") or entry["container_path"]
filename = f"{len(records):03d}-{entry['repository']}-{phase}.log"
started = time.time()
try:
with (output / filename).open("w") as log:
process = subprocess.run(["/bin/bash", "-o", "pipefail", "-c", text], cwd=working,
env=environment, stdout=log, stderr=subprocess.STDOUT, timeout=3600)
code = process.returncode
except (OSError, subprocess.TimeoutExpired) as error:
with (output / filename).open("a") as log:
log.write(str(error) + "\n")
code = 125
records.append({"repository": entry["repository"], "phase": phase, "exit_code": code,
"elapsed_seconds": round(time.time() - started, 3), "log": filename})
return code == 0
try:
if environment_setup and not command({"repository": "environment", "container_path": "/"},
"environment-setup", environment_setup):
infrastructure_failed = True
for _, entry in (source_order if not infrastructure_failed else []):
if not command(entry, "source-setup", entry.get("setup_command"), entry["container_path"]):
infrastructure_failed = True
break
if not infrastructure_failed:
for _, entry in ordered:
if not command(entry, "evaluation-setup", entry.get("evaluation_setup_command"), entry.get("evaluation_setup_cwd")):
infrastructure_failed = True
break
if not infrastructure_failed:
for _, entry in ordered:
if not command(entry, "ready", entry.get("evaluation_ready_command")):
infrastructure_failed = True
if not infrastructure_failed:
for phase in ("pass_to_pass", "fail_to_pass"):
for entry in commands:
if not command(entry, phase, entry.get(phase)):
tests_failed = True
finally:
for _, entry in reversed(ordered):
if not command(entry, "teardown", entry.get("evaluation_teardown_command")):
infrastructure_failed = True
summary = {"instance_id": contract["instance_id"], "evaluation_world": contract["evaluation_world"],
"infrastructure_failed": infrastructure_failed, "test_commands_failed": tests_failed, "verdict_scope": "Command execution only; native case inventories and task-specific acceptance rules must be checked separately.", "commands": records}
(output / "summary.json").write_text(json.dumps(summary, indent=2) + "\n")
# Exit codes alone are execution evidence, not a native case inventory verdict.
return 2 if infrastructure_failed else 1 if tests_failed else 0
if __name__ == "__main__":
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--contract", type=Path, required=True)
parser.add_argument("--output-dir", type=Path, required=True)
parser.add_argument("--environment-setup", help="Verified task environment preparation command; runs before frozen source setup")
args = parser.parse_args()
raise SystemExit(execute(json.loads(args.contract.read_text()), args.output_dir, args.environment_setup))
|