bluejay-rb
Warning This project is still very early in its development and should be considered highly unstable and experimental. It is incomplete and not ready for production use.
Note In an effort to get a working proof-of-concept, documentation and test coverage has been sacrificied. This will be corrected prior to a stable release.
bluejay-rb is a GraphQL engine for Ruby written primarily in Rust through the use of magnus.
Goal
bluejay-rb's goal is to provide a lightning fast GraphQL engine with first-class integration with Sorbet. If you do not care deeply about one or both of these goals, then you should seriously consider using the excellent graphql-ruby gem, which has many more features than bluejay-rb in addition to being production-ready and much more customizable.
Benchmarks
Some benchmarks comparing the performance of bluejay-rb against graphql-ruby (with C parser) are located at /bench. The results of some of these benchmarks are included below.
Profiling IPS:
Warming up --------------------------------------
graphql 158.000 i/100ms
bluejay 1.764k i/100ms
Calculating -------------------------------------
graphql 1.928k (± 2.7%) i/s - 9.638k in 5.003328s
bluejay 17.899k (± 2.7%) i/s - 89.964k in 5.029921s
Comparison:
bluejay: 17899.5 i/s
graphql: 1927.9 i/s - 9.28x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 45.944k memsize ( 2.736k retained)
421.000 objects ( 41.000 retained)
12.000 strings ( 11.000 retained)
bluejay 5.256k memsize ( 208.000 retained)
40.000 objects ( 2.000 retained)
0.000 strings ( 0.000 retained)
Comparison:
bluejay: 5256 allocated
graphql: 45944 allocated - 8.74x more
Profiling IPS:
Warming up --------------------------------------
graphql 79.000 i/100ms
bluejay 1.804k i/100ms
Calculating -------------------------------------
graphql 781.860 (± 2.2%) i/s - 3.950k in 5.054411s
bluejay 18.017k (± 1.6%) i/s - 90.200k in 5.007842s
Comparison:
bluejay: 18016.7 i/s
graphql: 781.9 i/s - 23.04x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 45.944k memsize ( 21.752k retained)
421.000 objects ( 197.000 retained)
12.000 strings ( 12.000 retained)
bluejay 5.256k memsize ( 5.056k retained)
40.000 objects ( 35.000 retained)
0.000 strings ( 0.000 retained)
Comparison:
bluejay: 5256 allocated
graphql: 45944 allocated - 8.74x more
Profiling IPS:
Warming up --------------------------------------
graphql 166.000 i/100ms
bluejay 1.101k i/100ms
Calculating -------------------------------------
graphql 1.642k (± 2.1%) i/s - 8.300k in 5.055691s
bluejay 10.876k (± 1.8%) i/s - 55.050k in 5.063054s
Comparison:
bluejay: 10876.3 i/s
graphql: 1642.4 i/s - 6.62x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 97.728k memsize ( 40.000 retained)
885.000 objects ( 1.000 retained)
6.000 strings ( 0.000 retained)
bluejay 15.640k memsize ( 80.000 retained)
228.000 objects ( 2.000 retained)
15.000 strings ( 1.000 retained)
Comparison:
bluejay: 15640 allocated
graphql: 97728 allocated - 6.25x more
Profiling IPS:
Warming up --------------------------------------
graphql 79.000 i/100ms
bluejay 1.009k i/100ms
Calculating -------------------------------------
graphql 841.882 (± 1.5%) i/s - 4.266k in 5.068484s
bluejay 10.163k (± 2.1%) i/s - 51.459k in 5.065688s
Comparison:
bluejay: 10163.2 i/s
graphql: 841.9 i/s - 12.07x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 97.688k memsize ( 30.736k retained)
884.000 objects ( 319.000 retained)
6.000 strings ( 5.000 retained)
bluejay 12.600k memsize ( 536.000 retained)
152.000 objects ( 6.000 retained)
8.000 strings ( 1.000 retained)
Comparison:
bluejay: 12600 allocated
graphql: 97688 allocated - 7.75x more
Profiling IPS:
Warming up --------------------------------------
bluejay 39.288k i/100ms
graphql 6.906k i/100ms
Calculating -------------------------------------
bluejay 397.916k (± 1.4%) i/s - 2.004M in 5.036417s
graphql 67.671k (± 4.9%) i/s - 338.394k in 5.015897s
Comparison:
bluejay: 397915.9 i/s
graphql: 67671.2 i/s - 5.88x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
bluejay 0.000 memsize ( 0.000 retained)
0.000 objects ( 0.000 retained)
0.000 strings ( 0.000 retained)
graphql 6.192k memsize ( 2.816k retained)
70.000 objects ( 37.000 retained)
6.000 strings ( 6.000 retained)
Comparison:
bluejay: 0 allocated
graphql: 6192 allocated - Infx more
Profiling IPS:
Warming up --------------------------------------
bluejay 37.618k i/100ms
graphql 6.034k i/100ms
Calculating -------------------------------------
bluejay 403.347k (± 2.1%) i/s - 2.031M in 5.038540s
graphql 59.928k (± 2.0%) i/s - 301.700k in 5.036461s
Comparison:
bluejay: 403347.5 i/s
graphql: 59928.3 i/s - 6.73x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
bluejay 0.000 memsize ( 0.000 retained)
0.000 objects ( 0.000 retained)
0.000 strings ( 0.000 retained)
graphql 6.192k memsize ( 0.000 retained)
70.000 objects ( 0.000 retained)
6.000 strings ( 0.000 retained)
Comparison:
bluejay: 0 allocated
graphql: 6192 allocated - Infx more
Profiling IPS:
Warming up --------------------------------------
bluejay 203.000 i/100ms
graphql 30.000 i/100ms
Calculating -------------------------------------
bluejay 2.106k (± 1.4%) i/s - 10.556k in 5.014156s
graphql 308.161 (± 1.9%) i/s - 1.560k in 5.064324s
Comparison:
bluejay: 2105.7 i/s
graphql: 308.2 i/s - 6.83x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
bluejay 0.000 memsize ( 0.000 retained)
0.000 objects ( 0.000 retained)
0.000 strings ( 0.000 retained)
graphql 1.425M memsize ( 556.448k retained)
16.001k objects ( 7.541k retained)
50.000 strings ( 50.000 retained)
Comparison:
bluejay: 0 allocated
graphql: 1425400 allocated - Infx more
Profiling IPS:
Warming up --------------------------------------
bluejay 208.000 i/100ms
graphql 27.000 i/100ms
Calculating -------------------------------------
bluejay 2.098k (± 1.6%) i/s - 10.608k in 5.057106s
graphql 274.835 (± 1.8%) i/s - 1.377k in 5.011880s
Comparison:
bluejay: 2098.2 i/s
graphql: 274.8 i/s - 7.63x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
bluejay 0.000 memsize ( 0.000 retained)
0.000 objects ( 0.000 retained)
0.000 strings ( 0.000 retained)
graphql 1.425M memsize ( 0.000 retained)
16.001k objects ( 0.000 retained)
50.000 strings ( 0.000 retained)
Comparison:
bluejay: 0 allocated
graphql: 1425400 allocated - Infx more
Profiling IPS:
Warming up --------------------------------------
graphql 504.000 i/100ms
bluejay 5.370k i/100ms
Calculating -------------------------------------
graphql 4.991k (± 6.2%) i/s - 25.200k in 5.073307s
bluejay 45.612k (±11.5%) i/s - 225.540k in 5.057627s
Comparison:
bluejay: 45612.5 i/s
graphql: 4991.1 i/s - 9.14x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 33.392k memsize ( 12.200k retained)
383.000 objects ( 152.000 retained)
17.000 strings ( 13.000 retained)
bluejay 40.000 memsize ( 40.000 retained)
1.000 objects ( 1.000 retained)
0.000 strings ( 0.000 retained)
Comparison:
bluejay: 40 allocated
graphql: 33392 allocated - 834.80x more
Profiling IPS:
Warming up --------------------------------------
graphql 299.000 i/100ms
bluejay 4.726k i/100ms
Calculating -------------------------------------
graphql 2.953k (± 5.0%) i/s - 14.950k in 5.075697s
bluejay 51.904k (± 4.0%) i/s - 259.930k in 5.016499s
Comparison:
bluejay: 51904.0 i/s
graphql: 2953.1 i/s - 17.58x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 33.304k memsize ( 12.320k retained)
381.000 objects ( 155.000 retained)
15.000 strings ( 13.000 retained)
bluejay 40.000 memsize ( 40.000 retained)
1.000 objects ( 1.000 retained)
0.000 strings ( 0.000 retained)
Comparison:
bluejay: 40 allocated
graphql: 33304 allocated - 832.60x more
Profiling IPS:
Warming up --------------------------------------
graphql 189.000 i/100ms
bluejay 4.228k i/100ms
Calculating -------------------------------------
graphql 1.929k (± 7.3%) i/s - 9.639k in 5.026688s
bluejay 43.632k (± 4.6%) i/s - 219.856k in 5.050598s
Comparison:
bluejay: 43631.9 i/s
graphql: 1929.2 i/s - 22.62x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 77.788k memsize ( 0.000 retained)
755.000 objects ( 0.000 retained)
50.000 strings ( 0.000 retained)
bluejay 699.000 memsize ( 0.000 retained)
1.000 objects ( 0.000 retained)
1.000 strings ( 0.000 retained)
Comparison:
bluejay: 699 allocated
graphql: 77788 allocated - 111.28x more
Profiling IPS:
Warming up --------------------------------------
graphql 157.000 i/100ms
bluejay 4.662k i/100ms
Calculating -------------------------------------
graphql 1.562k (± 1.7%) i/s - 7.850k in 5.027157s
bluejay 45.335k (± 3.3%) i/s - 228.438k in 5.044407s
Comparison:
bluejay: 45335.1 i/s
graphql: 1561.9 i/s - 29.02x slower
Profiling Ruby memory allocations:
Calculating -------------------------------------
graphql 77.884k memsize ( 675.000 retained)
755.000 objects ( 10.000 retained)
50.000 strings ( 7.000 retained)
bluejay 699.000 memsize ( 0.000 retained)
1.000 objects ( 0.000 retained)
1.000 strings ( 0.000 retained)
Comparison:
bluejay: 699 allocated
graphql: 77884 allocated - 111.42x more
To update the benchmark report above, run rake bench:doc.
Installation
Install the gem and add to the application's Gemfile by executing:
$ bundle add bluejay
If bundler is not being used to manage dependencies, install the gem by executing:
$ gem install bluejay
Usage
See an example in the /example directory. Note the usage of Tapioca RBI generation for the DSL at /example/sorbet/rbi/dsl/graph
Development
After checking out the repo, run bin/setup to install dependencies and rake compile to compile the Rust extension. Then, run rake test to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.
To install this gem onto your local machine, run bundle exec rake install.
To release a new version, update the version number in version.rb, and push that to main. Then, create a new release in the GitHub UI with a new tag equal to the version number in version.rb prefixed with v. This will trigger the release workflow which will publish the new version with pre-built binaries.