Descriptive part
Ontology design
From these definitions we deduce the classes and subclasses in yellow, the object properties in blue and the data properties in green. The details of the practical work will exceptionally be in French to coincide with the Protege software.
Classes
- Le beau temps et le mauvais temps sont deux types de phénomènes
- La pluie et le brouillard sont des types de phénomènes de mauvais temps, l’ensoleillement est un type de phénomène de beau temps
- Les paramètres mesurables sont une classe de concept, ainsi que les instants et les observations. La classe concept n'est pas présente sur la capture, il s'agit d'un oubli et ne gène pas la suite du TP
- Une ville, un pays et un continent sont des types de lieux
Object & data properties
- Un phénomène est caractérisé par des paramètres mesurables
- Un phénomène a une durée en minutes
- Un phénomène débute à un instant
- Un phénomène finit à un instant
- Un instant a un timestamp, de type xsd :dateTimeStamp
- Un phénomène a pour symptôme une observation
- Une observation météo mesure un paramètre mesurable
- Une observation météo a une valeur pour laquelle vous ne représenterez pas l’unité
- Une observation météo a pour localisation un lieu.
- Une observation météo a pour date un instant
- Un lieu peut être inclus dans un autre lieu ▶ est situé dans
- Un lieu peut inclure un autre lieu ▶ contient
- Un pays a pour capitale une ville
The second step is to populate this ontology with several individuals following certain rules. The details will not be present here but you can find them in my report. "Protege" incorporates a tool called hermit reasoner that allows you to make inferences about individuals. For example, if we create Paris as an untyped individual and France as a country, we can then create the triplet France :a pour capitale Paris. The reasoner then deduces that Paris is a city. This is a very simple example that shows how it works. It uses the established rules to deduce additional information.
Exploiting the ontology
The goal of this second part is to use the previously created ontology and to implement to implement functions that will be used by this ontology. We then use the properties defined in part 1 to represent some individuals of the ontology.
The code:
main functions
At first, the objective is to implement the IModelFunctions interface using the functions present in IConvenienceInterface. So we have five functions to to implement. Each function corresponds to a jUnit test that we must validate.
This function takes as parameter the name of a Place and returns the URI of the created individual. Its goal is to create an instance of the Location class with the name of the parameter given as argument of the function.
The purpose of this function is the same as createPlace and takes as input a TimestampEntity as an instant is characterized by its timestamp. However, this new individual is created only if there is no other instant associated with the same timestamp.
This function returns the URI of the Instant according to the timestamp given in parameter if it exists.
This function returns the timestamp associated to the Instant instance passed in parameter.
The purpose of this function is the same as createPlace and takes as input a TimestampEntity as an instant is characterized by its timestamp. However, this new individual is created only if there is no other instant associated with the same timestamp.
This function allows you to create an observation from the different elements that characterize it its value, its "Instant" and the type of the parameter that is observed. The function function returns the URI of the created observation. An observation being more complex, it is therefore longer than the function to longer than the function to create a Place.
This function parses the list of observations extracted from the dataset and adds them to our knowledge base. It takes as argument the list of observations and the URI of the phenomenon.