blob: 38b04286b516b9e60eb50f339de228adb0cbf0c5 [file] [edit]
#!/usr/bin/env python3
# Copyright 2026 The Dawn & Tint Authors
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
# 1. Redistributions of source code must retain the above copyright notice, this
# list of conditions and the following disclaimer.
#
# 2. Redistributions in binary form must reproduce the above copyright notice,
# this list of conditions and the following disclaimer in the documentation
# and/or other materials provided with the distribution.
#
# 3. Neither the name of the copyright holder nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
"""Merge script to aggregate LiteRT-LM benchmark results and upload metrics to GCS.
"""
import json
import os
from pathlib import Path
import subprocess
import sys
import tempfile
import time
DAWN_ROOT = Path(__file__).resolve().parent.parent.parent
sys.path.insert(0, str(DAWN_ROOT))
from testing.merge_scripts import merge_api
def find_gsutil() -> list[str]:
gsutil_path = DAWN_ROOT / 'third_party' / 'depot_tools' / 'gsutil.py'
if not gsutil_path.exists():
raise RuntimeError(f'Unable to find gsutil.py at {gsutil_path}')
return [sys.executable, '-u', str(gsutil_path)]
def upload_file_to_gcs(local_file_path: Path, gcs_bucket: str,
gcs_dest_path: str) -> None:
if not local_file_path.exists():
raise FileNotFoundError(f"Local file not found at: {local_file_path}")
gsutil_cmd = find_gsutil()
gcs_url = f"gs://{gcs_bucket}/{gcs_dest_path}"
cmd = gsutil_cmd + ['cp', str(local_file_path), gcs_url]
print(f"Uploading {local_file_path.name} to GCS: {' '.join(cmd)}")
subprocess.run(cmd, check=True)
def generate_metadata(metadata_file_path: Path, timestamp: int,
build_properties_str: str | None) -> None:
props = {}
if build_properties_str:
try:
props = json.loads(build_properties_str)
except Exception as e:
print(f"Warning: Failed to parse build-properties: {e}",
file=sys.stderr)
metadata = {
'timestamp': timestamp,
'git_revision': props.get('got_revision'),
'buildername': props.get('buildername'),
'builder_group': props.get('builder_group'),
}
with open(metadata_file_path, 'w', encoding='utf-8') as f:
json.dump(metadata, f, indent=2)
def main() -> int:
parser = merge_api.ArgumentParser()
args = parser.parse_args()
# This benchmark is designed to run only on a single shard.
# Raise an error if more than one shard output is passed to be merged.
if len(args.jsons_to_merge) > 1:
raise ValueError(
f"Expected exactly 1 shard for litert_lm_benchmark, "
f"but found {len(args.jsons_to_merge)} shards: {args.jsons_to_merge}"
)
# Merge standard test result JSONs (for exactly 0 or 1 shard).
merged_results = {
'failures': [],
'valid': True,
}
for json_file_str in args.jsons_to_merge:
json_file = Path(json_file_str)
if not json_file.exists():
continue
try:
with open(json_file, 'r', encoding='utf-8') as f:
results = json.load(f)
if results.get('failures'):
merged_results['failures'].extend(results['failures'])
if not results.get('valid', True):
merged_results['valid'] = False
except Exception as e:
print(f"Error reading {json_file}: {e}", file=sys.stderr)
merged_results['valid'] = False
# Write the combined results JSON required by the merge API.
output_json_path = Path(args.output_json)
with open(output_json_path, 'w', encoding='utf-8') as f:
json.dump(merged_results, f, indent=2)
# Locate litert_lm_metrics.pb from the single shard output.
metric_file = None
if args.jsons_to_merge:
json_file = Path(args.jsons_to_merge[0])
isolated_outdir = json_file.parent
pb_path = isolated_outdir / 'litert_lm_metrics.pb'
if pb_path.exists():
metric_file = pb_path
if not metric_file:
print("Warning: No litert_lm_metrics.pb file found to upload.",
file=sys.stderr)
return 0
# Upload metrics and generated metadata to GCS.
timestamp = int(time.time())
task_id = os.environ.get('SWARMING_TASK_ID', 'local')
run_id = f"{timestamp}_{task_id}"
bucket_name = 'dawn-webgpu-perf-results'
directory_name = f'litert_lm_benchmark/{run_id}'
with tempfile.TemporaryDirectory() as tempdir_str:
tempdir = Path(tempdir_str)
# Generate metadata.json locally with accurate builder context.
metadata_file = tempdir / 'metadata.json'
generate_metadata(metadata_file, timestamp,
getattr(args, 'build_properties', None))
upload_file_to_gcs(metadata_file, bucket_name,
f"{directory_name}/metadata.json")
# Upload the single metric file.
upload_file_to_gcs(metric_file, bucket_name,
f"{directory_name}/litert_lm_metrics.pb")
return 0
if __name__ == '__main__':
sys.exit(main())