def test_impart_taxons():
# Imparting a knowledge decreases the learning cost of the corresponding taxons, but has no
# effect whatsoever on the selected programs.
dbf = ProgramFilter(db)
dbf.impart_taxons({"flow/conditional"})
print(set(dbf.selected_programs.keys()))
assert set(dbf.selected_programs.keys()) == {
"assignment.py",
"collatz_print.py",
"fizzbuzz.py",
"is_even.py",
}
def test_impart_taxons():
dbf = ProgramFilter(db)
taxons = ["O/J", "X/S/M/L", "non/existing/taxon"]
dbf.impart_taxons(taxons)
print(sorted(dbf.imparted_knowledge))
assert dbf.imparted_knowledge == {
"O",
"O/J",
"X",
"X/S",
"X/S/M",
"X/S/M/L",
"non",
"non/existing",
"non/existing/taxon",
}
def test_impart_taxons():
# Imparting a triple doesn't make much sense. Currently, it comes down to imparting the two
# taxons, and ignoring the predicate.
dbf = ProgramFilter(db)
dbf.impart_taxons({("equals", "operator/arithmetic/modulo", "type/number/integer")})
print(set(dbf.selected_programs.keys()))
assert set(dbf.selected_programs.keys()) == {
"assignment.py",
"collatz_print.py",
"fizzbuzz.py",
"is_even.py",
}
print(dbf.imparted_knowledge)
assert dbf.imparted_knowledge == {
"operator",
"operator/arithmetic",
"operator/arithmetic/modulo",
"type",
"type/number",
"type/number/integer",
}
