Allow pre-compilation of rules conditions

[luispfgarces]

Description

Adds functionality to pre-compile rules' conditions in order to speed up evaluation times. Putting it simply: it adds a middleware component that intercepts rules loaded from the data source or being written to the data source, and compiles them to a lambda expression. The resultant lambda expression is stored on the data source, thanks to the new IConditionNode<TConditionType>.Properties dictionary property, which enables to store the compiled lambda expression. This compiled lambda expression is then loaded from the data source avoiding re-compiling the rules' conditions again. This development also includes a conditions eval engine prepared to evaluate compiled rules.

A new project with a Benchmark.Net implementation was already added to measure some performance numbers. Unfortunately, I was unable to assess the improvement of this implementation vs. the master codebase. I'm including the performance numbers taken from the implementation as-is for your consideration.

Benchmark	Method	EnableCompilation	Mean (ns)	Error (ns)	StdDev (ns)	Skewness	Kurtosis	BranchInstructions/Op	BranchMispredictions/Op	Gen0	Allocated (KB)
1 - Test rule (1 rule, 1 condition)	RunAsync	FALSE	660,5	3,08	2,88	-0,0707	1,751	2122	2	0,1287	1,59
1 - Test rule (1 rule, 1 condition)	RunAsync	TRUE	584,1	3,69	3,27	-0,0777	2,724	1841	2	0,1402	1,72
2 - Songs (2 rules, various conditions)	RunAsync	FALSE	2003,9	9,34	8,73	-0,3514	2,111	7089	8	0,3166	3,92
2 - Songs (2 rules, various conditions)	RunAsync	TRUE	958,8	5,35	5,01	-0,2135	2,135	3619	2	0,1564	1,93
3 - Poker rules (100+ rules, various conditions)	RunAsync	FALSE	39190	263	246	-0,166	2,221	141307	239	6,2866	77,37
3 - Poker rules (100+ rules, various conditions)	RunAsync	TRUE	20910	148	139	0,676	2,303	75102	65	2,9297	36,15

Here's how they compare:

Benchmark	Time taken [Interpreted] - ns	Time taken [Compiled] - ns	Time taken [Interpreted vs. Compiled] - % gain	Memory [Interpreted] - KB	Memory [Compiled] - KB	Memory [Interpreted vs. Compiled] - % gain
1	660,50	584,1	11,57%	1,59	1,72	-8,18%
2	2003,9	958,8	52,15%	3,92	1,93	50,77%
3	39190	20910	46,64%	77,37	36,15	53,28%

Related to #49.

Change checklist

• Code follows the code rules guidelines of this project
• Commit messages follow the commit rules of this project
• I have self-reviewed my changes before submitting this pull request
• I have covered new/changed code with new tests and/or adjusted existent ones
• I have made changes necessary to update the documentation accordingly -> will adjust the wiki after merge to master.

Please also check the I want to contribute guidelines and make sure you have done accordingly.

Disclaimer

By sending us your contributions, you are agreeing that your contribution is made subject to the terms of our Contributor Ownership Statement

[jdromeiro]

More of a nit-pick, but in the header of the Pull Request consider changing Classic to Interpreted.

[jdromeiro]

👏🏼👏🏼👏🏼