Unit:
Department of Informatics and TelecommunicationsΠληροφορική
Author:
POLYTSERIS ZACHARIAS
Supervisors info:
Εμμανουήλ Κουμπαράκης, Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών.
Κωνσταντίνος Πλας, Βοηθός Ερευνητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών.
Original Title:
Nomothesi@: Migrating the Linked Data into Elasticsearch
Translated title:
Nomothesi@: Migrating the Linked Data into Elasticsearch
Summary:
The main objective of this thesis was the back-end redevelopment of the Greek Legislation Platform, Nomothesi@ and an evaluation performance test between the two implementations. The platform used a triple store for legislative RDF data which got replaced with an Elasticsearch document store in order to scale up, accelerate queries and provide new features and functionality that take advantage of Elasticsearches’ engine. The most important step to achieve that was the transformation of the Turtle datasets into structured JSON documents along with the creation of the new indexes’ context definition. The entirety of the code regarding indexing and searching data had to be reformatted by replacing the SPARQL queries with the elastic equivalents to maintain existing functionality, resulting in a cleaner Nomothesi@ back-end that supports full-text searches, aggregations and relevance scores.
Main subject category:
Technology - Computer science
Keywords:
Elasticsearch, Indexing, RDF to ndJSON, Full-Text Search, Performance Testing, REST Services, Linked Open Data
Elastic_Nomothesia_web_platform.pdf (382 KB)
Open in new window
additional_tools.zip
3 KB
File access is restricted.