More performance improvement to log density evaluation

[sunxd3]

This remove another performance bottleneck: getindex from MetaGraph. This is done by storing the data in arrays.

The size of a BUGSModel is increased by 15%.

And it roughly half the previous time, while not impacting the ReverseDiff time.

[coveralls]

Pull Request Test Coverage Report for Build 11541635140

Details

• 50 of 51 (98.04%) changed or added relevant lines in 3 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.4%) to 78.246%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/model.jl	46	47	97.87%
<!--	Total:	50	51	98.04%	-->

Totals
Change from base Build 11519174461:	0.4%
Covered Lines:	1410
Relevant Lines:	1802

---

💛 - Coveralls

[github-actions[bot]]

BridgeStan not found at location specified by $BRIDGESTAN environment variable, downloading version 2.5.0 to /home/runner/.bridgestan/bridgestan-2.5.0 Done! Model dogs produces error: ErrorException("log_density() failed with exception: Exception: bernoulli_lpmf: Probability parameter is inf, but must be in the interval [0, 1] (in '/home/runner/work/JuliaBUGS.jl/JuliaBUGS.jl/benchmark/stan/bugs_examples/vol1/dogs/dogs.stan', line 37, column 6 to line 38, column 62)\n")	Model	Parameter Count	Data Count	Stan Density Time (µs)	Stan Density Gradient Time (µs)	JuliaBUGS Density Time with Graph Walk (µs)	JuliaBUGS Density Gradient Time with ReverseDiff.jl(compiled tape) (µs)
rats	65	150	5.5624	8.07167	138.207	85.5195
pumps	12	10	0.978833	1.28504	18.825	6.34925
dogs	2	720	NA	NA	265.154	150.305
seeds	26	21	2.51291	3.12922	49.1165	21.53
surgical_realistic	14	12	1.26496	1.61911	24.715	8.506
magnesium	108	96	10.3645	11.9825	237.342	73.848
salm	22	18	2.37017	3.10306	39.283	12.383
equiv	15	20	2.50373	3.39375	27.351	17.984
dyes	9	30	1.02618	1.33386	16.716	12.964
stacks	6	21	1.12596	1.68841	52.619	14.647
epil	303	236	32.821	38.231	603.015	239.336
blockers	47	44	3.28944	3.73063	107.992	28.283
oxford	244	240	16.13	18.726	704.884	173.784
lsat	1006	5000	173.574	214.736	4083.15	1187.36
bones	33	422	73.758	91.48	1356.48	190.204
mice	20	65	7.32375	9.451	42.279	37.48
kidney	64	58	10.765	16.261	121.757	76.332
leuk	18	714	20.428	26.9	336.116	205.393
leukfr	40	714	23.804	31.949	371.893	239.897

[sunxd3]

This is not perfect, but I am merging this now and will optimize memory size later.