TeMex: The Web Template Extractor

This paper presents and describes TeMex, a site-level web template extractor. TeMex is fully automatic, and it can work with online webpages without any preprocessing stage (no information about the template or the associated webpages is needed) and, more importantly, it does not need a predefined set of webpages to perform the analysis. TeMex only needs a URL. Contrarily to previous approaches, it includes a mechanism to identify webpage candidates that share the same template. This mechanism increases both recall and precision, and it also reduces the amount of webpages loaded and processed. We describe the tool and its internal architecture, and we present the results of its empirical evaluation.