From fc13ce44781b86511ef5b90ac1ad8e836df7a401 Mon Sep 17 00:00:00 2001 From: theoreticalbts Date: Wed, 21 Oct 2015 13:27:40 -0400 Subject: [PATCH] genesis_util: Implement create_bloom_filter --- programs/genesis_util/change_key_prefix.py | 63 +++++++++++++++++ programs/genesis_util/create_bloom_filter.py | 72 ++++++++++++++++++++ 2 files changed, 135 insertions(+) create mode 100755 programs/genesis_util/change_key_prefix.py create mode 100755 programs/genesis_util/create_bloom_filter.py diff --git a/programs/genesis_util/change_key_prefix.py b/programs/genesis_util/change_key_prefix.py new file mode 100755 index 00000000..7ca4b876 --- /dev/null +++ b/programs/genesis_util/change_key_prefix.py @@ -0,0 +1,63 @@ +#!/usr/bin/env python3 + +import argparse +import json +import sys + +def dump_json(obj, out, pretty): + if pretty: + json.dump(obj, out, indent=2, sort_keys=True) + else: + json.dump(obj, out, separators=(",", ":"), sort_keys=True) + return + +def main(): + parser = argparse.ArgumentParser(description="Set is_prefixed=false for all asset owners") + parser.add_argument("-o", "--output", metavar="OUT", default="-", help="output filename (default: stdout)") + parser.add_argument("-i", "--input", metavar="IN", default="-", help="input filename (default: stdin)") + parser.add_argument("-f", "--from", metavar="PREFIX", default="", help="initial prefix") + parser.add_argument("-t", "--to", metavar="PREFIX", default="", help="new prefix") + parser.add_argument("-p", "--pretty", action="store_true", default=False, help="pretty print output") + opts = parser.parse_args() + + if opts.input == "-": + genesis = json.load(sys.stdin) + else: + with open(opts.input, "r") as f: + genesis = json.load(f) + + frum = opts.__dict__["from"] # from is a language keyword and cannot be an attribute name + + def convert(k): + if not k.startswith(frum): + print("key {k} does not start with prefix {p}".format(k=repr(k), p=repr(frum))) + raise RuntimeError() + return opts.to + k[len(frum):] + + for account in genesis["initial_accounts"]: + account["owner_key"] = convert(account["owner_key"]) + account["active_key"] = convert(account["active_key"]) + + for asset in genesis["initial_assets"]: + for cr in asset.get("collateral_records", []): + cr["owner"] = convert(cr["owner"]) + + for balance in genesis["initial_balances"]: + balance["owner"] = convert(balance["owner"]) + + for vb in genesis["initial_vesting_balances"]: + vb["owner"] = convert(vb["owner"]) + + for witness in genesis["initial_witness_candidates"]: + witness["block_signing_key"] = convert(witness["block_signing_key"]) + + if opts.output == "-": + dump_json( genesis, sys.stdout, opts.pretty ) + sys.stdout.flush() + else: + with open(opts.output, "w") as f: + dump_json( genesis, f, opts.pretty ) + return + +if __name__ == "__main__": + main() diff --git a/programs/genesis_util/create_bloom_filter.py b/programs/genesis_util/create_bloom_filter.py new file mode 100755 index 00000000..17576d35 --- /dev/null +++ b/programs/genesis_util/create_bloom_filter.py @@ -0,0 +1,72 @@ +#!/usr/bin/env python3 + +import argparse +import hashlib +import json +import sys + +def main(): + parser = argparse.ArgumentParser(description="Set is_prefixed=false for all asset owners") + parser.add_argument("-o", "--output", metavar="OUT", default="-", help="output filename (default: stdout)") + parser.add_argument("-i", "--input", metavar="IN", default="-", help="input filename (default: stdin)") + parser.add_argument("-n", "--num", metavar="N", default=3, type=int, help="number of hashes per key (default: 3)") + parser.add_argument("-s", "--size", metavar="BITS", default=8*1048576, type=int, help="number of bits in filter") + parser.add_argument("-a", "--algorithm", metavar="NAME", default="sha256", type=str, help="hash algorithm (must exist in hashlib)") + opts = parser.parse_args() + + if opts.input == "-": + genesis = json.load(sys.stdin) + else: + with open(opts.input, "r") as f: + genesis = json.load(f) + + keys = set() + + for account in genesis["initial_accounts"]: + keys.add(account["owner_key"]) + keys.add(account["active_key"]) + + for asset in genesis["initial_assets"]: + for cr in asset.get("collateral_records", []): + keys.add(cr["owner"]) + + for balance in genesis["initial_balances"]: + keys.add(balance["owner"]) + + for vb in genesis["initial_vesting_balances"]: + keys.add(vb["owner"]) + + for witness in genesis["initial_witness_candidates"]: + keys.add(witness["block_signing_key"]) + + sys.stderr.write("got {n} distinct keys\n".format(n=len(keys))) + + keys = [(str(i) + ":" + k).encode("UTF-8") for k in sorted(keys) for i in range(opts.num)] + + data = bytearray((opts.size + 7) >> 3) + + h = getattr(hashlib, opts.algorithm) + for k in keys: + address = int(h(k).hexdigest(), 16) % opts.size + data[address >> 3] |= (1 << (address & 7)) + + popcount = [bin(i).count("1") for i in range(256)] + w = sum(popcount[x] for x in data) + sys.stderr.write("""w={w} o={o:.3%} p={p:.3%} +w: Hamming weight o: Occupancy p: False positive probability +""".format( +w=w, +o=float(w) / float(opts.size), +p=(float(w) / float(opts.size))**opts.num, +)) + + if opts.output == "-": + sys.stdout.write(data) + sys.stdout.flush() + else: + with open(opts.output, "wb") as f: + f.write(data) + return + +if __name__ == "__main__": + main()