CARTRAWLER CODE ASSESSMENT

Starting from the provided source, below are the changes introduced:

• The original cars were saved into a CSV file which is then loaded by using the apache-commons-csv library (introduced in the pom) - loading from an external source introduces flexibility, and improves testability.
• The list of cars was loaded, as the original code, into a Set, although a list would be a more natural container. Deduping on a list is a bit more involved, thus for simplicity the Set was kept as it's a structure that provide the dedupe functionality out of the box. Although Collection was used as input/output parameter, as they are a better abstraction and would allow changing Set to List with no breaking of the downstream processing.
• To perform check for duplicates the CarResult implements an equals and hashCode method that uses only the field required for the deduplication as required (i.e. the price is not used in both methods); equals and hashCode were implemented using helper methods from Apache commons-lang3. Although this may not implement the "natural" equality, those, together with loading data into a Set, allowed "for free" the resulting deduplication of the input dataset.
• A separate class takes care of the sorting according to spec, as per SOLID pattern
• Filtering of the above median FULFULL cars is handled by a separate service (again, SOLID). Behaviour is togglable via command line by specifying the --filterAboveMedian switch.
• Introduction of command line allowed to use also a different input file for testing.
• Maven pom is set up so it packages everything into a runnable jar that can be invoked directly, like

  java -jar challenge-1.0-SNAPSHOT-jar-with-dependencies.jar [--filterAboveMedian] [file to read]