fabiolimace
commented
3 years ago Created codecs for base-36, base-58 and base-62.
Also created static methods for codecs of any radix between 2 and 64.
List of changes:
Add Base36Codec Add Base58Codec Add Base58BitcoinCodec // Bitcoin's base-58 Add Base58FlickrCodec // Flickr's base-58 Add Base62Codec Add BaseNCodec.getInstance() // static factory for any radix between 2 and 64 Add test cases for base-16, base-32 and base-64 with 3k samples Add test cases for base-36, base-58 and base-62 with 3k samples
Test coverage: 90.0%
Benchmark for base-n codecs:
------------------------------------------------------------------
Benchmark Mode Cnt Score Error Units
------------------------------------------------------------------
Throughput.decode_base16 thrpt 5 4695,920 ± 6,835 ops/ms
Throughput.decode_base32 thrpt 5 2146,940 ± 26,668 ops/ms
Throughput.decode_base64 thrpt 5 2566,964 ± 25,757 ops/ms
Throughput.decode_base36 thrpt 5 528,938 ± 3,441 ops/ms *
Throughput.decode_base58 thrpt 5 609,171 ± 8,800 ops/ms *
Throughput.decode_base62 thrpt 5 629,332 ± 2,297 ops/ms *
------------------------------------------------------------------
Throughput.encode_base16 thrpt 5 18432,630 ± 171,769 ops/ms
Throughput.encode_base32 thrpt 5 20345,201 ± 126,980 ops/ms
Throughput.encode_base64 thrpt 5 21841,282 ± 43,678 ops/ms
Throughput.encode_base36 thrpt 5 267,618 ± 22,657 ops/ms *
Throughput.encode_base58 thrpt 5 243,444 ± 20,533 ops/ms *
Throughput.encode_base62 thrpt 5 257,363 ± 14,720 ops/ms *
------------------------------------------------------------------
(*) The new codecs don't perform as well as the others because the new
ones do integer arithmetic on BigInteger objects.
List of new alphabets (char sets):
The alphabet for base-58 is the same used in Bitcoin.