• Scala 2.12.4
• sbt 1.0.3
• Java 8

cd into the project folder and execute sbt test

• README.md
• build.sbt
• src - main - scala - SearchApp.scala
• src - test - resources
• src - test - scala - FinderSpec.scala
• src - test - scala - SearcherSpec.scala

In main, SearchApp.scala contains

• package object, which is used to define tiny types to represent the data structure used in the application.
• SearchApp, which is the entry point of the application, and holds files data instantiated during start up in a mutable variable.
• Searcher, the main class called by SearchApp to perform startup (data initialisation), search (see below) and output result.
• Finder, the class Searcher used to delegate tasks to traverse file system and load the content of the files.

In test, there are resources folder containing fixtures for testing, and two specs files for testing the two main classes.

There are the following parts of the search algorithm

• Finding all possible combinations of the input value

  • The input value (string) is first being split into a list of strings (determined by space).
  • Then find out all combinations of the words (formed by the words next to each other), such as "The quick brown fox" has 10 possible combinations.

• Finding matching keywords in the content

  • First find out if there is a match on the first word from the keywords in the content, keywords is a subset (one or more words) from the input keywords.
  • If there is a match, then recursively find the existence of the second word in the content, so on and so forth.
  • If there is no match, either the search will stop and return the matching result, or return empty result.

• Finding all possible combinations of the input value (keywords) in all the loaded contents

• Rank the result by using the formula: matchingKeywordsFound / totalNumberOfKeywordsFromInput

• Produce the final result