Replace get! with get to avoid unnecessary initialization of state values

[bowenszhu]

The original code used get! which would initialize state[t] to 0 if it didn't exist before incrementing. This is unnecessary as the value is only used for incrementing. Using get with a default value of 0 achieves the same result without potential performance overhead.

See the source code for get:
https://github.com/JuliaLang/julia/blob/0b4590a5507d3f3046e5bafc007cacbbfc9b310b/base/dict.jl#L545
and for get!:
https://github.com/JuliaLang/julia/blob/0b4590a5507d3f3046e5bafc007cacbbfc9b310b/base/dict.jl#L473

[github-actions]

Benchmark Results

	master	5320dd7...	master/5320dd74f75ffc...
overhead/acrule/a+2	0.753 ± 0.021 μs	0.725 ± 0.021 μs	1.04
overhead/acrule/a+2+b	0.746 ± 0.019 μs	0.721 ± 0.021 μs	1.04
overhead/acrule/a+b	0.254 ± 0.012 μs	0.262 ± 0.013 μs	0.97
overhead/acrule/noop:Int	25.4 ± 0.92 ns	26.2 ± 0.93 ns	0.967
overhead/acrule/noop:Sym	0.0334 ± 0.0059 μs	0.0342 ± 0.0051 μs	0.976
overhead/rule/noop:Int	0.0379 ± 0.00043 μs	0.0376 ± 0.001 μs	1.01
overhead/rule/noop:Sym	0.0439 ± 0.0015 μs	0.0424 ± 0.0013 μs	1.04
overhead/rule/noop:Term	0.0435 ± 0.0016 μs	0.0428 ± 0.0018 μs	1.02
overhead/ruleset/noop:Int	0.127 ± 0.0036 μs	0.127 ± 0.0021 μs	1
overhead/ruleset/noop:Sym	0.133 ± 0.0037 μs	0.131 ± 0.0028 μs	1.01
overhead/ruleset/noop:Term	6.75 ± 0.54 μs	6.59 ± 0.42 μs	1.02
overhead/simplify/noop:Int	0.141 ± 0.0033 μs	0.14 ± 0.0027 μs	1.01
overhead/simplify/noop:Sym	0.151 ± 0.0038 μs	0.154 ± 0.0039 μs	0.981
overhead/simplify/noop:Term	0.0444 ± 0.0023 ms	0.045 ± 0.0027 ms	0.988
overhead/simplify/randterm (+, *):serial	0.133 ± 0.0032 s	0.135 ± 0.0019 s	0.984
overhead/simplify/randterm (+, *):thread	0.0868 ± 0.027 s	0.0842 ± 0.028 s	1.03
overhead/simplify/randterm (/, *):serial	0.272 ± 0.0088 ms	0.279 ± 0.0093 ms	0.973
overhead/simplify/randterm (/, *):thread	0.324 ± 0.01 ms	0.334 ± 0.011 ms	0.97
overhead/substitute/a	0.121 ± 0.0028 ms	0.122 ± 0.0031 ms	0.992
overhead/substitute/a,b	0.0995 ± 0.0027 ms	0.103 ± 0.0027 ms	0.966
overhead/substitute/a,b,c	16.7 ± 0.69 μs	19.5 ± 0.86 μs	0.856
polyform/easy_iszero	0.053 ± 0.0025 ms	0.0527 ± 0.0021 ms	1
polyform/isone	2.79 ± 0.01 ns	2.79 ± 0.01 ns	1
polyform/iszero	5.57 ± 0.1 ms	5.54 ± 0.081 ms	1
polyform/simplify_fractions	3.38 ± 0.065 ms	3.39 ± 0.048 ms	0.997
time_to_load	4.56 ± 0.048 s	4.57 ± 0.0037 s	0.997

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).