There are two classes of tests (for now): query and construct, each in its own directory. The structure is identical.

Each test is in a different python file (usually Test***.py, where *** refer to the chapter
in the SPARQL document). The test is run by the queryTest (resp. constructTest) script.
This script does a run-time import of the test file, getting the variables in the Test files,
construct and runs the sparql query and displays the result.

Each test file has an identical structure: it is a bunch of (python) declaration:

- the rdf data itself ("rdfData")
	it is also possible to refer to an array of external files using the 
	variable "datafiles", but the rdfData takes precedence. If files are used,
	either set rdfData to None, or do not set it at all
- the 'where' and 'optional' clauses, set as GraphPatterns
- the 'select' tuple (if applicable)
- the 'construct' pattern (if applicable)
- for the query case, the expected result should be added in the form of a string

Looking at the current examples, the structure is pretty straightforward.