"""Compare serial/parallel throughput on the same four validation configurations."""
import hashlib
import json
from pathlib import Path
import subprocess
import sys

root = Path('/home/int/run/cats-dogs')
out = root / 'acceleration'
out.mkdir(exist_ok=True)
script = Path(__file__).with_name('run_batch.py')
summaries = {}
for mode, workers in [('serial', 1), ('parallel', 4)]:
    jobs = []
    for threshold in (600, 1600, 4000, 8000):
        jobs.append({'name': f'bench_{mode}_t{threshold}', 'args': [
            '--data-dir', str(root / 'features-v2'), '--dataset', 'sparse_hog_validation',
            '--train', '4000', '--count', '1000', '--threshold', str(threshold),
            '--columns', '20', '--quant', 'weights', '--backend', 'single']})
    plan, summary = out / f'{mode}-plan.json', out / f'{mode}-summary.json'
    plan.write_text(json.dumps({'jobs': jobs}, indent=2) + '\n')
    subprocess.run([sys.executable, str(script), str(plan), '--jobs', str(workers),
                    '--summary', str(summary)], check=True)
    summaries[mode] = json.loads(summary.read_text())
checks = {}
for threshold in (600, 1600, 4000, 8000):
    checks[str(threshold)] = {}
    for file in ('float_weights.csv', 'quant_weights.csv', 'float_neurons.csv', 'quant_neurons.csv', 'result.csv'):
        a = (root / f'bench_serial_t{threshold}' / file).read_bytes()
        b = (root / f'bench_parallel_t{threshold}' / file).read_bytes()
        checks[str(threshold)][file] = a == b
assert all(value for files in checks.values() for value in files.values()), 'Parallel run changed results'
result = {'serial': summaries['serial'], 'parallel': summaries['parallel'],
          'speedup': summaries['serial']['elapsed_seconds'] / summaries['parallel']['elapsed_seconds'],
          'bitwise_equal': checks}
(out / 'batch-benchmark.json').write_text(json.dumps(result, indent=2) + '\n')
print('SPEEDUP', result['speedup'], flush=True)