#!/usr/bin/env python3
"""Small synthetic size experiment. No throughput or production-ranking claims."""
from pathlib import Path
from collections import Counter
import hashlib, json, math, random, platform, sys
import zstandard as zstd
import brotli
ROOT=Path(__file__).resolve().parents[1]
N=262144
rng=random.Random(20261011)
records=[]
while sum(map(len,records))<N:
 i=len(records)
 records.append((json.dumps({'sensor':i%32,'tick':i,'temperature':round(18+(i%100)/10,1),'state':['idle','active','active','idle'][i%4]},separators=(',',':'))+'\n').encode())
inputs={'periodic':bytes(range(256))*(N//256),'telemetry':b''.join(records)[:N],'pseudorandom':rng.randbytes(N)}
encoders={
 'zstd-3':(lambda b:zstd.ZstdCompressor(level=3).compress(b),lambda b:zstd.ZstdDecompressor().decompress(b)),
 'zstd-19':(lambda b:zstd.ZstdCompressor(level=19).compress(b),lambda b:zstd.ZstdDecompressor().decompress(b)),
 'brotli-5':(lambda b:brotli.compress(b,quality=5,mode=brotli.MODE_GENERIC,lgwin=22),brotli.decompress),
 'brotli-11':(lambda b:brotli.compress(b,quality=11,mode=brotli.MODE_GENERIC,lgwin=22),brotli.decompress)
}
report={'purpose':'Illustrative synthetic size experiment, not a general ranking','date':'2026-10-11','input_bytes_each':N,'python':sys.version.split()[0],'platform':platform.platform(),'zstandard_binding':zstd.__version__,'libzstd':list(zstd.ZSTD_VERSION),'brotli':brotli.__version__,'dictionary':'none trained; Brotli format static dictionary is implicit in decoder','sizes_include':'complete returned compressed streams; shared codec software excluded','samples':{}}
for name,data in inputs.items():
 d={'sha256':hashlib.sha256(data).hexdigest(),'raw_bytes':len(data),'byte_histogram_entropy':-sum((c/N)*math.log2(c/N) for c in Counter(data).values()),'compressed':{}}
 for label,(encode,decode) in encoders.items():
  blob=encode(data);assert decode(blob)==data
  d['compressed'][label]={'bytes':len(blob),'percent_of_input':100*len(blob)/N,'round_trip_exact':True,'sha256':hashlib.sha256(blob).hexdigest()}
 report['samples'][name]=d
(ROOT/'evidence/measurements.json').write_text(json.dumps(report,indent=2)+'\n')
print(json.dumps(report,indent=2))