test-task: Advanced Rule Matcher

Description

There is set of rules with following structure: RuleId, Rule Priority, Filter1.. Filter4.

RuleId	Priority	Filter1	Filter2	Filter3	Filter4
1	80	AAA	<ANY>	CCC	DDD
2	10	<ANY>	<ANY>	AAA	<ANY>
3	70	BBB	<ANY>	CCC	<ANY>
4	100	AAA	BBB	CCC	<ANY>
5	50	CCC	AAA	<ANY>	CCC
6	0	<ANY>	<ANY>	<ANY>	<ANY>

Purpose of the library to find best matching rules for datasets with filtering values (Filter1, ..., Filter4).

• If record matches multiple rules, rule with highest priority must be selected (100 – highest priority, 0 – lowest priority).
• Value <ANY> means, that any value is accepted for this filter.

Initially rules are stored in CSV file.

In real production environment quantity of rules expected to be 1000+. Quantity of data to pass through filter is 10x-100x larger or might be even infinity data stream in future.

Must have features:

• To load data from a file to array of strongly typed entities
• To create a library that will be finding best matching rule (respecting rule priority) for the set of input values
• To unit tests the library

Sample to describe idea:

Rule? rule = searchEngine.FindRule(string val1, string val2, string val3, string val4);

Expected, that this module will be used to find matching rules with the frequency 100 calls/second over loaded dataset.

Should have features:

• Taking into account that collection can store 1000+ records it should be performant enough.
• Code should be reusable as much as possible to implement different Rules models. For instance, to write another set of strongly typed rules with filters: Filter1(int), Filter2(bool), Filter3(string).

Validation

In case of provided sample data (see a table in Description or CSV file), expected results will be following:

Value1	Value2	Value3	Value4		RuleId
AAA	BBB	CCC	AAA		4
AAA	BBB	CCC	DDD		4
AAA	AAA	AAA	AAA		2
BBB	BBB	BBB	BBB		6
BBB	CCC	CCC	CCC		3