Thanks to the ever-increasing use of the Semantic Web, a growing number of entities (e.g. documents) are characterized by non-ambiguous mean-ings. Based on this characterization, entities can subsequently be compared us-ing semantic measures. A plethora of measures have been designed given their critical importance in numerous treatments relying on ontologies. However, the improvement and use of semantic measures are currently hampered by the lack of a dedicated theoretical framework and an extensive generic software solution dedicated to them. To meet these needs, this paper presents a unified theoretical framework of graph-based semantic measures, from which we developed the open source Semantic Measures Library and toolkit; a solution that paves the way for straightforward design, computation and analysis of semantic measures for both users and developers. Downloads, documentation and technical support at dedicated website http://www.semantic-measures-library.org.