Upload DiscoGeM 2.0

column_explain.txt

@@ -0,0 +1,30 @@
+split		
+train/dev/test splits
+
+genre
+genre of the text: Europarl or Novel
+
+doc
+source document. Europarl proceeding ID or name of the novel
+
+orig_lang
+the original language of the text; other languages are translation
+
+available_langs
+list of language versions available for that item
+
+{MV|WAWA|ZBS|GLAD|MACE|DS}_{en|de|fr|cs}
+single aggregated label by each of the 7 aggregation methods for each language version of the item; empty values mean the language version is not included in that item
+
+{MV|WAWA|ZBS|GLAD|MACE|DS}_dist_{en|de|fr|cs}
+dictionary of the label distribution obtained by each aggregation method for each language version of the item; empty values mean the language version is not included in that item
+
+dist20_{MV|WAWA|ZBS|GLAD|MACE|DS}_{en|de|fr|cs}
+list of labels with distribution >= 0.2 based on the distribution obtained by each aggregation method for each language version of the item; empty values mean the language version is not available, but  [] means none of the labels have >= 0.2 distribution; 0.2 was chosen as the threshold based on evaluation against a subset of labels with reference annotation in DiscoGeM 1.0
+
+norm_entropy_{en|de|fr|cs}
+normalized entropy of the original label distribution (i.e. MV_dist) for each language version; NONE means the language version is not available
+
+{Arg1/arg2}_context_{en|de|fr|cs}
+The two text spans between which the annotators annotate the connective
+

rel_conn_maps.csv

@@ -0,0 +1,30 @@
+level3	level2	level1	en_conn	de_conn	fr_conn	cs_conn
+synchronous	synchronous	temporal	at the same time	gleichzeitig	en même temps	zároveň
+precedence	asynchronous	temporal	then	dann	ensuite	potom
+succession	asynchronous	temporal	after	davor,	après que	předtím
+reason	cause	contingency	bause	weil	parce que	protože
+result	cause	contingency	as a result	daher	c'est pourquoi	proto
+arg1-as-goal	purpose	contingency	for that purpose	dazu	à cette fin	za tím účelem
+arg2-as-goal	purpose	contingency	so that	sodass	afin que	aby
+arg1-as-cond	condition	contingency	in that case	insofern	dans ce cas	v tom případě
+arg1-as-negcond	condition	contingency	if not	sonst	sinon	jinak
+arg2-as-cond	condition	contingency	if	wenn	si	pokud
+arg2-as-negcond	condition	contingency	unless	es sei denn	à moins que	ledaže
+arg1-as-denier	concession	comparison	even though	obwohl	bien que	a to i přesto, že
+arg2-as-denier	concession	comparison	nonetheless	trotzdem	néanmoins	přesto
+contrast	contrast	comparison	on the other hand	andererseits	d'autre part	na druhou stranu
+similarity	similarity	comparison	similarly	gleichermaßen	de même	podobně
+equivalence	equivalence	expansion	in other words	anders gesagt	en d'autres termes	jinými slovy
+arg1-as-instance	instantiation	expansion	this illustrates that 	das verdeutlicht, dass 	cela illustre que	to je příkladem toho, že
+arg2-as-instance	instantiation	expansion	for example	zum Beispiel	par exemple	například
+arg1-as-detail	level-of-detail	expansion	in short	um es kurz zu machen	bref	zkrátka
+arg2-as-detail	level-of-detail	expansion	in more detail	genauer gesagt	plus précisement	konkrétně
+conjunction	conjunction	expansion	also	darüberhinaus	en plus	také
+disjunction	disjunction	expansion	or	oder	ou	nebo
+arg1-as-excpt	exception	expansion	other than that	abgesehen von dieser Ausnahme	à part cela/ça	kromě této výjimky
+arg2-as-excpt	exception	expansion	an exception is that	eine Ausnahme ist, dass	une exception est que	výjimkou je to, že
+arg1-as-manner	manner	expansion	thereby	hiermit	de cette manière	tímto způsobem
+arg2-as-manner	manner	expansion	as if	mittels	comme si	následujícím způsobem:
+arg1-as-subst	substitution	expansion	rather than	anstatt, dass	plutôt que	místo, aby
+arg2-as-subst	substitution	expansion	instead	stattdessen	au lieu de	místo toho
+norel	norel	norel	(no direct relations)	(keine direkte Beziehung)	(pas de relation directe)	(bez přímého vztahu)