feat: add sample benchmarks

[curtisy1]

Adds a benchmark accessing translations in 1m key value pair files. Results are mostly the same but could be a bit inaccurate since a few duplicates exist which resx might handle better than us

Method	Mean	Error	StdDev
FirstEntryJson	64.97 ns	0.790 ns	0.700 ns
FirstEntryResx	65.11 ns	0.460 ns	0.408 ns
LastEntryJson	69.01 ns	1.451 ns	1.727 ns
LastEntryResx	63.58 ns	0.845 ns	0.790 ns
MiddleEntryJson	67.91 ns	0.567 ns	0.474 ns
MiddleEntryResx	67.24 ns	1.004 ns	0.838 ns
SameEntryTwiceJson	126.66 ns	1.731 ns	1.620 ns
SameEntryTwiceResx	138.74 ns	0.926 ns	0.821 ns