Closed JarbasAI closed 7 years ago
Initial Tests seem to be working:
output:
joana can be: wife
joana is: human animal mammal
Code
name = "human"
child_concepts = ["male", "female", "name"]
parent_concepts = ["animal", "mammal"]
knowledge.create_concept(name, parent_concepts=parent_concepts, child_concepts=child_concepts)
name = "joana"
child_concepts = ["wife"]
parent_concepts = ["human"]
knowledge.create_concept(name, parent_concepts=parent_concepts, child_concepts=child_concepts)
name = "animal"
child_concepts = ["dog", "cow", "frog", "cat", "spider", "insect"]
parent_concepts = ["alive"]
knowledge.create_concept(name, parent_concepts=parent_concepts, child_concepts=child_concepts)
# lets see what concept connector can deduce from here
key = "joana"
childs = knowledge.concepts[key].child_concepts
parents = knowledge.concepts[key].parent_concepts
Some structure will be needed to keep track of current concepts and update connections
this structure must:
creating a new concept with childs needs to:
creating a new concept with parents need to:
question: parents of parents could also be updated with new concept as child (if human is a mammal and mammal is an animal, human is also an animal), how far up should this go?
Work started in commit https://github.com/HeinzSchmidt/LILACS-mycroft-core/commit/add04885018546589a200e292360b4840bf117d7