Webis EditorialSum Corpus 2020

The Webis EditorialSum Corpus consists of 1330 manually curated extractive summaries for 266 news editorials spanning three diverse portals: Al-Jazeera, Guardian and Fox News. Each editorial has 5 summaries, each labeled for overall quality and fine grained properties such as thesis-relevance, persuasiveness, reasonableness, self-containedness. The files are organized as follows: <br> <em>corpus.csv</em> - <strong>Contains all the editorials and their acquired summaries</strong> <br> Note: (X = [1,5] for five summaries) - article_id : Article ID in the corpus<br> - title : Title of the editorial<br> - article_text : Plain text of the editorial<br> - summary_{X}_text : Plain text of the corresponding summary<br> - thesis_{X}_text : Plain text of the thesis from the corresponding summary<br> - lead : top 15% of the editorial's segments<br> - body : segments between lead and conclusion sections<br> - conclusion : bottom 15% of the editorial's segments<br> - article_segments: Collection of paragraphs, each further divided into collection of segments containing:<br> { "number": segment order in the editorial,<br> "text" : segment text,<br> "label": ADU type<br> }<br> - summary_{X}_segments: Collection of summary segments containing:<br> { "number": segment order in the editorial,<br> "text" : segment text,<br> "adu_label": ADU type from the editorial,<br> "summary_label": can be 'thesis' or 'justification'<br> } <br> <em>quality-groups.csv</em> - <strong>Contains the IDs for high(and low)-quality summaries for each quality dimension per editorial</strong><br> <br> For example: article_id 2 has four high_quality summaries (summary_1, summary_2, summary_3, summary_4) and one low_quality summary (summary_5) in terms of overall quality.<br> The summary texts can be obtained from corpus.csv respectively.