platform_build_soong/bloaty/bloaty_merger.py

80 lines
2.5 KiB
Python
Raw Normal View History

# Copyright 2021 Google Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Bloaty CSV Merger
Merges a list of .csv files from Bloaty into a protobuf. It takes the list as
a first argument and the output as second. For instance:
$ bloaty_merger binary_sizes.lst binary_sizes.pb
"""
import argparse
import csv
import ninja_rsp
import file_sections_pb2
BLOATY_EXTENSION = ".bloaty.csv"
def parse_csv(path):
"""Parses a Bloaty-generated CSV file into a protobuf.
Args:
path: The filepath to the CSV file, relative to $ANDROID_TOP.
Returns:
A file_sections_pb2.File if the file was found; None otherwise.
"""
file_proto = None
with open(path, newline='') as csv_file:
file_proto = file_sections_pb2.File()
if path.endswith(BLOATY_EXTENSION):
file_proto.path = path[:-len(BLOATY_EXTENSION)]
section_reader = csv.DictReader(csv_file)
for row in section_reader:
section = file_proto.sections.add()
section.name = row["sections"]
section.vm_size = int(row["vmsize"])
section.file_size = int(row["filesize"])
return file_proto
def create_file_size_metrics(input_list, output_proto):
"""Creates a FileSizeMetrics proto from a list of CSV files.
Args:
input_list: The path to the file which contains the list of CSV files. Each
filepath is separated by a space.
output_proto: The path for the output protobuf.
"""
metrics = file_sections_pb2.FileSizeMetrics()
reader = ninja_rsp.NinjaRspFileReader(input_list)
for csv_path in reader:
file_proto = parse_csv(csv_path)
if file_proto:
metrics.files.append(file_proto)
with open(output_proto, "wb") as output:
output.write(metrics.SerializeToString())
def main():
parser = argparse.ArgumentParser()
parser.add_argument("input_list_file", help="List of bloaty csv files.")
parser.add_argument("output_proto", help="Output proto.")
args = parser.parse_args()
create_file_size_metrics(args.input_list_file, args.output_proto)
if __name__ == '__main__':
main()