different samples treated as replicates

[RobFryer]

resolves #512

Throws an error if there are any missing values in the sample column of the contaminants file. (Only likely to happen with external data.)

Also checks for non-unique sample identifiers, for example when the same sample identifier has been used in different years, or for different stations or species. If this occurs, then a warning is printed exhorting the user to sort out their contaminants data file. But if they don't an attempt is made to create unique sample identifiers by pasting together year, station_code, species (biota only) and sample. This should work in most situations.

Have updated the vignette on the external data file format to point out the need for unique sample identifiers (across years, stations, species, etc.)

Tested on HELCOM and external data