A repo to test the compression ratio (not performance!) of JavaScript client side string compression for encoding in the URL. The flow looks like this:
string --> compress(string) --> toBase64(compressedString)
The ratio is then measured by the length of the compressed base64 string. To run the benchmarks, run yarn benchmark. To run tests, run yarn test.
| no compression | Base64 | ULZSS | LZString | LZW | GZIP | ZLIB | DEFLATE | |
|---|---|---|---|---|---|---|---|---|
| sketch | 1320 | 1760 | 1276 | 1100 | 1404 | 864 | 848 | 840 |
| simplex noise | 2936 | 3916 | 2768 | 2244 | 2968 | 1708 | 1692 | 1684 |
| quil | 7745 | 10328 | 5572 | 4444 | 6096 | 3196 | 3180 | 3172 |
| triangle | 9112 | 12152 | 5224 | 4876 | 6936 | 2808 | 2792 | 2784 |
| modele | 23766 | 31688 | 15980 | 12292 | 19100 | 7776 | 7760 | 7752 |
| no compression | Base64 | u-lzss | lz-string | LZW | GZIP | ZLIB | DEFLATE | |
|---|---|---|---|---|---|---|---|---|
| sketch | 100% | 133.33% | 96.66% | 83.33% | 106.36% | 65.45% | 64.24% | 63.63% |
| simplex noise | 100% | 133.37% | 94.27% | 76.43% | 101.08% | 58.17% | 57.62% | 57.35% |
| quil | 100% | 133.35% | 71.94% | 57.37% | 78.70% | 41.26% | 41.05% | 40.95% |
| triangle | 100% | 133.33% | 57.33% | 53.51% | 76.11% | 30.81% | 30.64% | 30.55% |
| modele | 100% | 133.33% | 67.23% | 51.72% | 80.36% | 32.71% | 32.65% | 32.61% |
The u-lzss algorithm by Google is buggy! With large strings, the second half of the string becomes malformed. It passes only 2/3 tests.