Defining new concepts¶
This chapter presents how to define new concepts. You can link defined entities and create a new knowledge. This is represented by the NEW_DEFINITION token. This token is very similar to ENTITY introduced in the previous chapter. However, while ENTITY must be defined in the Concepts definition, this does not hold for NEW_DEFINITION. Moreover, NEW_DEFINITION name can be followed by a proposition. This is the case of a verb, e.g. work in, assigned to… As propositions, the tool currently accepts: “to”, “by”, “in”.
Note that as said above, since it is a new knowledge, it is possible to not define these concepts in the concepts definition. If the new concept is not defined, the structure is computed as follows:
all the keys of the entities defined as a condition are added to the new definition;
you can define specific attributes with the “with ATTRIBUTE_NAME VALUE”;
if an attribute is already included it is discarded.
Simple definition¶
The simplest way to define a new concept is:
SIMPLE_DEFINITION -> ENTITY AUXILIARY_VERB? NEW_DEFINITION LIST_OF_ENTITIES?.
where AUXILIARY_VERB can be one of the verbs to be, to have, and to do and the optional LIST_OF_ENTITIES is a comma separated list of ENTITY.
Examples:
An author has written a movie. (NEW_DEFINITION = written)Waiter Alice is working in a pub. (NEW_DEFINITION = working_in)Director Spielberg has written a movie with name equal to JurassicPark. (NEW_DEFINITION = written)Note the usage of the proposition in in the second sentence.
Corresponding ASP:¶
Considering the following definitions:
An author is identified by a name.
A waiter is identified by a name.
A director is identified by a name.
A movie is identified by a name.
The previous examples are translated into:
written(MOVIDE_NAME,AUTHOR_NAME) :- movie(MOVIDE_NAME), author(AUTHOR_NAME).
working_in("Alice",PB_NM) :- waiter("Alice"), pub(PB_NM).
written("spielberg","jurassicPark") :- movie("jurassicPark"), author("spielberg").
Note that in both sentences, written has been defined by the tool taking the keys of the entities in the body.
Quantified choice proposition¶
A quantified choice proposition is defined as follows:
QUANTIFIED_CHOICE_PROPOSITION -> ("Every" | "Any") ENTITY ASSIGNMENT_VERB AUXILIARY_VERB? NEW_DEFINITION CARDINALITY? LIST_OF_ENTITIES?.
where ASSIGNMENT_VERB is one of “can” and “must”. The difference between “can” and “must”, is that with “must” whenever the condition is satisfied there is the NEW_DEFINITION, instead, with can whenever the condition is satisfied you can have or not the NEW_DEFINITION (this corresponds to the ASP assignment and ASP choice rules). Moreover, with can, you can define a CARDINALITY:
CARDINALITY -> ("exactly" | "at most" | "at least") STRING
| "between" STRING "and" STRING
Finally, LIST_OF_*ENTITIES* is a comma-separated list of ENTITY.
Examples:
Every patron can drink in exactly 1 pub.
Every waiter can serve a drink.
Every movie with name I must have a scoreAssignment with movie I, with value 5.
Corresponding ASP:¶
Considering the following definitions:
A patron is identified by a name.
A pub is identified by a name.
A waiter is identified by a name.
A drink is identified by a name.
The previous example is translated into:
1 <= {drink_in(PATRON_NAME,PUB_NAME): pub(PUB_NAME)} <= 1 :- patron(PATRON_NAME).
{serve(WAITER_NAME,DRINK_NAME): drink(DRINK_NAME)} :- waiter(WAITER_NAME).
scoreassignment(I,5) :- movie(I).
Note that the scoreAssignment structure has been explicitly declared using the with ATTRIBUTE_NAME VALUE statement.
Whenever/Then propositions¶
First of all, a WHENEVER_CLAUSE is defined as follows:
WHENEVER_CLAUSE -> "Whenever there is" NEGATION? ENTITY
This proposition allows you to define a condition (in ASP, it adds elements to a rule body), and NEGATION is optional and is equal to “not” allowing to negate the condition.
A Whenever-then proposition is:
WHENEVER_THAN_CLAUSE -> LIST_OF_WHENEVER_CLAUSE "then" ENTITY ASSIGNMENT_VERB AUXILIARY_VERB? [CARDINALITY] NEW_DEFINITION LIST_OF_ENTITIES
where, as previously, LIST_OF_ENTITIES is a comma separated list of ENTITY.
Examples:
Whenever there is a movie with director equal to spielberg, with id X then we must have a topmovie with id X. (in this example we have an ENTITY = *we*, it is possibile to use pronouns and in that case, it is ignored and not converted into any entity)
Whenever there is a movie M, then M can be assigned to exactly 1 director.
The whenever there is is used to define the condition, that, if satisfied, defines the new concept.
Corresponding ASP:¶
Considering the following definitions:
A movie is identified by an id, and has a name.
A director is identified by a name.
The previous example is translated into:
topmovie(X) :- movie(X,"JurassikPark").
1 <= {assigned_to(M,DRCTR_NM): director(DRCTR_NM)} <= 1 :- movie(M,_).