# Some examples

To play with an example, open it with Cogui
• either download the cogxml file from this page and open it as a file
• or paste the link in the URL tab of the Cogui open dialog box
 examplePaperICCS2010.cogxml (in English and French) This is the basic example used in the paper published at ICCS 2010: "Translations between RDF(S) and Conceptual Graphs" by Baget, Croitoru, Gutierrez, Leclere and Mugnier. finalExampleOfTutorial.cogxml (in English) This is the final example of the tutorial. It contains facts, queries and rules. childhood.cogxml (in English) This is a very simple introductory example (i.e. without nestings, explicit coreference links, modules, patterns …) dealing with common sense knowledge concerning children playing. Some peculiarities of this example: The concept type hierarchy is a partial order but is not a tree. The relation hierarchy contains unary, binary and ternary relations. Relations of any arity greater than 0 can be considered. The graph G1.2 contains a cycle (it represents the following fact « a girl and a child are playing with the same toy, and the girl is a relative of the boy »). A conceptual graph may contain any number of cycles, cf. the graph 2.5 which is more complex than G1.2. H2.4 contains parallel edges (the relation vertex corresponds to a reflexive verb). A graph is not necessarily connected, e.g. H1.3 is disconnected. Let us consider the «query facts» task. Using G1.2 as a query (you can copy/paste from the fact to the query wizard) and ticking only K1.2 into the set_1 of facts you see that there is a projection (homomorphism) from G1.2 to K1.2. It means that K1.2 is an answer to G1.2 (equivalently, K1.2 is a specialization of G1.2 and G1.2 is a generalization of K1.2) . There is also a projection from H1.3 to K1.2. H2.9 and K2.9 are also two specializations of G2.9. Cogui compute all projections from the query graph to the fact graphs. There are two projections from G2.10 to H2.10. Note that a projection is not necessarily injective (i.e., one-to-one) nor surjective. The set of rules contains three rules. Try a query search in set_1 with the query Q1 in set_2 with and without using the rules. There is no answer without the rules and there is one answer with rule1. Bucolic.cogxml (in English and French) This small example illustrates nested graphs: «a person is thinking of the painting A, which represents a bucolic scene (first nesting), which can be described as follows: there is a couple in a boat on a lake (second nesting); this couple is composed (third nesting) of a person who is fishing truts in this lake and of a person who is sleeping (and this person is the same as the person who is thinking of the painting)». More examples to come...

SudocAD is a system dedicated to the author linkage problem in the French national bibliographic infrastructure context.
SudocAD is the result of a collaboration between ABES (http://www.abes.fr), the French National Agency in charge of this infrastructure and GraphIK (http://www.lirmm.fr/graphik).

The goal of SudocAD is to solve the following author linkage problem:
Given a bibliographic record about a document d in the Persee bibliographic base (http://www.persee.fr/web/guest/home), for each author in d SudocAD computes a qualitatively ordered list of authorities in the sudoc authority base which could be linked to this author. This ordered result can be used in different ways, allowing different strategies for the linkage problem (in an automatic mode as well as in a decision-aided mode). The file NoticePerseeExample.rdfs is the Persee record used in this presentation of SudocAD.

You can browse through the ontology which contains a hierarchy of (313) concepts and a hierarchy of (1179) relations. The root of the linkage relations is liage : binding vocabulary relation (Resource,Resource).
The script "X_COMPLET / EXAMPLE_GUY_FAURE" allows to understand how the system works. It is composed of seven steps briefly described as follows (Information about currently running process is given on the java console).
• STEP1 representation of a Persee bibliographic record into the GBKR language.
• STEP2 import of the Sudoc data obtained from an author in the Persee record. This step has been inhibited because it uses web services accessing the sudoc catalog and the authority base IdRef that contains the sudoc catalog (http://www.abes.fr/Sudoc/Sudoc-public) and the authority base (http://en.abes.fr/Other-services/IdRef). The resulting working graph is Facts/rdf-imports/sudoc_guy_faure
• STEP3 for each author candidate in sudoc_guy_faure graph, the information used in the linkage procedure are computed (domain profile, publication interval, language list).
• STEP4 for each candidate, the comparison criteria are computed for the attributes: denomination, domain, date, language.
• STEP5 the 22 rules used in sudocAD are fired (in a specific order in such a way that their semantics is equivalent to 300 logical rules).
• STEP6 and 7 construction of the file containing the results and cleaning of working space.