-
Notifications
You must be signed in to change notification settings - Fork 1
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
2 changed files
with
169 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,142 @@ | ||
"""Script for benchmarking different message compression options.""" | ||
|
||
import argparse | ||
import json | ||
|
||
TITLE_SPACING = 20 | ||
SIZE_SPACING = 8 | ||
|
||
|
||
def compare_size(algo_pkg, data): | ||
"""Compare the size of the data in various compression methods.""" | ||
|
||
was_bytes = False | ||
if isinstance(data, bytes): | ||
was_bytes = True | ||
print(f"{'raw bytes':{TITLE_SPACING}} {len(data):{SIZE_SPACING}}") | ||
import base64 | ||
|
||
data = base64.b64encode(data).decode("utf-8") | ||
print(f"\nthe following use base64 first...\n") | ||
|
||
compressed_size = len(algo_pkg.compress(json.dumps(data).encode())) | ||
print( | ||
f"{f'json+encode+{algo_pkg.__name__}':{TITLE_SPACING}} {compressed_size:{SIZE_SPACING}}" | ||
) | ||
|
||
json_encoded_size = len(json.dumps(data).encode()) | ||
print( | ||
f"{'only json+encode':{TITLE_SPACING}} {json_encoded_size:{SIZE_SPACING}} (comp. ratio: {compressed_size / json_encoded_size:.3f})" | ||
) | ||
|
||
import pickle | ||
|
||
if not was_bytes: | ||
pickle_size = len(pickle.dumps(data)) | ||
print( | ||
f"{'only pickled':{TITLE_SPACING}} {pickle_size:{SIZE_SPACING}} (comp. ratio: {compressed_size / pickle_size:.3f})" | ||
) | ||
|
||
|
||
def compress_decompress(algo_pkg, data): | ||
if isinstance(data, bytes): | ||
import base64 | ||
|
||
data = base64.b64encode(data).decode("utf-8") | ||
comp_data = algo_pkg.compress(json.dumps(data).encode()) | ||
out_data = json.loads(algo_pkg.decompress(comp_data)) | ||
# just in case there's a typo, check the data | ||
assert out_data == data | ||
|
||
|
||
def main(): | ||
parser = argparse.ArgumentParser() | ||
parser.add_argument("--algo", required=True) | ||
parser.add_argument("--data", required=True) | ||
parser.add_argument("--only-size", action="store_true") | ||
args = parser.parse_args() | ||
|
||
# get data | ||
|
||
if args.data == "large_dict": | ||
data = { | ||
chr(i): [{chr(k % (i + 1)): k} for k in range(1000)] for i in range(100) | ||
} | ||
elif args.data == "medium_string": | ||
data = "\n".join( | ||
[ | ||
"The quick brown fox jumps over the lazy dog.", | ||
"She sells sea shells by the sea shore.", | ||
"How much wood would a woodchuck chuck if a woodchuck could chuck wood?", | ||
"Peter Piper picked a peck of pickled peppers.", | ||
"A journey of a thousand miles begins with a single step.", | ||
"To be or not to be, that is the question.", | ||
"All that glitters is not gold.", | ||
"A picture is worth a thousand words.", | ||
"Actions speak louder than words.", | ||
"Better late than never.", | ||
"Birds of a feather flock together.", | ||
"Fortune favors the bold.", | ||
"A watched pot never boils.", | ||
"Beauty is in the eye of the beholder.", | ||
"Beggars can't be choosers.", | ||
"Cleanliness is next to godliness.", | ||
"Discretion is the better part of valor.", | ||
"Every cloud has a silver lining.", | ||
"Good things come to those who wait.", | ||
"Haste makes waste.", | ||
"If it ain't broke, don't fix it.", | ||
"Laughter is the best medicine.", | ||
"Necessity is the mother of invention.", | ||
"No man is an island.", | ||
"Practice makes perfect.", | ||
"Rome wasn't built in a day.", | ||
"The early bird catches the worm.", | ||
"The pen is mightier than the sword.", | ||
] | ||
) | ||
elif args.data == "skyscan_i3_frame_pkl": | ||
data = b"\x80\x04\x95\x96\x02\x00\x00\x00\x00\x00\x00\x8c\x10icecube._icetray\x94\x8c\x07I3Frame\x94\x93\x94)R\x94}\x94Bh\x02\x00\x00[i3]\x06\x00\x00\x00\x00\x00P\x05\x00\x00\x00\r\x00\x00\x00I3EventHeader\r\x00\x00\x00I3EventHeader\x86\x00\x00\x00\x00\x01\x02\x00\r\x00\x00\x00I3EventHeader\x01\x03\x00\x00\x00\x00\x01\x00\x01\x00\x00\x00>\x16\x02\x00\x00\x00\x00\x00\x94\x88t\x00\x14\x00\x00\x00\x08\x00\x00\x00#\x00\x00\x00SCAN_nside0001_pixel0008_posvar0000\x01\x00\x02\x00\x00\x00\x03\x00\x00\x00\xe6\x07\x00\x00\xc1\x9d\xde^\xe8,\x18\x02\x04\x00\x00\x00\x05\x00\x00\x00\xe6\x07\x00\x00o\xa0\xe1^\xe8,\x18\x02\x15\x00\x00\x00MillipedeSeedParticle\n\x00\x00\x00I3Particle\x96\x00\x00\x00\x00\x01\x02\x00\n\x00\x00\x00I3Particle\x01\x05\x00\x00\x00\x00\x01\x00\x01\x00\x00\x00\x14\x00\x00\x00N\t\xb7q\xd0\xd0\x14U\x00\x00\x00\x00(\x00\x00\x00\x00\x00\x00\x00\x01\x00\x02\x00\x00\x00\x03\x00\x00\x00\xaeG\xe1z\x14\xdec@H\xe1z\x14\xae\x93u\xc0\n\xd7\xa3p=\x86\x7f\xc0\x01\x00\x04\x00\x00\x00\x05\x00\x00\x00\xa8aF\x06fj\x02@\x88\xe5:\xd1\xc7\x07\x0c@\x00\x00\x00\x00jr&A\x00\x00\x00\x00\x00\x00\xf8\x7f\x00\x00\x00\x00\x00\x00\xf8\x7f\x14\xa3\xac\xb4\xcc/\xd3?\x00\x00\x00\x00\x11\x00\x00\x00SCAN_HealpixNSide\x10\x00\x00\x00I3PODHolder<int>\x1d\x00\x00\x00\x00\x01\x02\x00\x05\x00\x00\x00I3Int\x01\x00\x00\x00\x00\x00\x01\x00\x01\x00\x00\x00\x01\x00\x00\x00\x11\x00\x00\x00SCAN_HealpixPixel\x10\x00\x00\x00I3PODHolder<int>\x1d\x00\x00\x00\x00\x01\x02\x00\x05\x00\x00\x00I3Int\x01\x00\x00\x00\x00\x00\x01\x00\x01\x00\x00\x00\x08\x00\x00\x00\x1b\x00\x00\x00SCAN_PositionVariationIndex\x10\x00\x00\x00I3PODHolder<int>\x1d\x00\x00\x00\x00\x01\x02\x00\x05\x00\x00\x00I3Int\x01\x00\x00\x00\x00\x00\x01\x00\x01\x00\x00\x00\x00\x00\x00\x00\x1aU\x07\xa1\x94\x86\x94b." | ||
else: | ||
raise ValueError("invalid --data value") | ||
|
||
# go! | ||
|
||
if args.algo == "bz2": | ||
import bz2 | ||
|
||
if args.only_size: | ||
compare_size(bz2, data) | ||
else: | ||
compress_decompress(bz2, data) | ||
|
||
elif args.algo == "lzma": | ||
import lzma | ||
|
||
if args.only_size: | ||
compare_size(lzma, data) | ||
else: | ||
compress_decompress(lzma, data) | ||
|
||
elif args.algo == "zstd": | ||
import zstd # type: ignore | ||
|
||
if args.only_size: | ||
compare_size(zstd, data) | ||
else: | ||
compress_decompress(zstd, data) | ||
|
||
elif args.algo == "gzip": | ||
import gzip | ||
|
||
if args.only_size: | ||
compare_size(gzip, data) | ||
else: | ||
compress_decompress(gzip, data) | ||
|
||
else: | ||
raise Exception("invalid compression algorithm specified") | ||
|
||
|
||
if __name__ == "__main__": | ||
main() |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,27 @@ | ||
#!/bin/bash | ||
set -e | ||
|
||
if [ -z $1 ]; then | ||
echo "MISSING ARG: message_compression_benchmarking.sh N_ITERATIONS" | ||
exit 1 | ||
fi | ||
N_ITERATIONS=$1 | ||
|
||
all_algos='bz2 lzma zstd gzip' | ||
all_data='large_dict medium_string skyscan_i3_frame_pkl' | ||
|
||
for algo in $all_algos; do | ||
echo | ||
echo "==============================================================" | ||
echo "$algo:" | ||
for data in $all_data; do | ||
echo | ||
echo "--------------------------------------" | ||
echo "data=$data" | ||
echo | ||
python3 message_compression_benchmarking.py --algo $algo --data $data --only-size | ||
echo | ||
echo "iterations=$N_ITERATIONS (compress + decompress)" | ||
time for i in $(seq 1 $N_ITERATIONS); do python3 message_compression_benchmarking.py --algo $algo --data $data; done | ||
done | ||
done |