SemRep.v1.8_XML_output_desc.txt

Title: SemRep XML Output option
Author: Halil Kilicoglu

SemRep -X option SemRep generates XML output, which is more verbose than the standard output 
format and almost identical to full fielded output format content-wise.

The data elements for XML output are as follows:

-|SemRepAnnotation|: The top level XML element that contains a number of |Document|s and SemRep
output for each |Document|.

-|Document|: An individual document processed with SemRep that contains a number of |Utterance|s.
Its attributes are id (00000000, if not explicitly stated in input) and text, which contains
the full extent of the |Document| text.

-|Utterance|: A single sentence with its associated MetaMap and SemRep annotations. 
Its attributes are |id|, |section|, |type|, |number| and |text|. |id| attribute is an identifier for the sentence,
|section| indicates the section abbreviation (current options are 'ti' for title, 'ab' for abstract,
and 'tx' for everything else), |type| (section header, such as RESULTS, METHODS), |number| indicates 
its position in the section and |text| indicates the sentence string.

-|Entity|: An instance of a concept mapping. The attributes are |id|, |cui| (a unique UMLS concept
identifier or any other unique source identifier), |name| (a preferred name for the concept),
|semtypes| (semantic types associated with the concept), |text| (the textual mention for the concept),
|score| (MetaMap score for the mapping, 0-1000), |negated| (whether the concept is negated in text or not, true/false), |begin| (character offset for the first character of
the textual mention, 0-based), |end| (character offset for the end of the mention). Some attributes are relevant
only for genomic concepts (genes, proteins, etc.). These attributes are |entrezID| (a unique identifier from
the EntrezGene for the corresponding term), |entrezName| (the name of the corresponding term from
the EntrezGene).

-|Predication|: An instance of a SemRep relation that consists of |Subject|, |Predicate| and |Object| elements.
Its attributes are |id|, |negated| (whether the predication is negated or not, true/false), and |inferred| (whether
the predication is the result of an inference based on other predications).

-|Coreference|: An instance of an anaphora relation that consists of an |Anaphor| and an |Antecedent| element.
Its only attribute is |id|. 

-|Scale|: This element indicates a scale that is associated with a comparative |Predication|. Its attributes
are |id|, |name| (the name of the scale, such as 'effectiveness'), |text| (the textual indicator of the scale),
|begin| (character offset of the first character of the scale indicator), |end| (character offset of the last 
character of the scale indicator).

-|Subject|/|Object|: These elements correspond to the subject and object arguments of the parent |Predication|.
Their attributes are |entityID| (the identifier for the entity argument --a reference to an existing |Entity| element), 
|relSemType| (the semantic type of  the entity that is used for the predication), |dist| (distance in terms of 
noun phrases between the argument and the predicate), |maxDist| (maximum possible distance in terms of 
noun phrases from the predicate in the direction of the argument). Note that |dist| and |maxDist| are both zero,
if they are not relevant or not computed. 

-|Anaphor|/|Antecedent|: These elements correspond to the arguments of the parent |Coreference|.
Their single attribute is |entityID| (the identifier for the entity argument --a reference to an existing |Entity| element). 

-|Predicate|: The predicate that indicates the |Predication|. Its attributes are |type| (the ontological
predicate, such as TREATS, INHIBITS, etc.), |indicatorType| (the part-of-speech, syntactic structure, or
mechanism that indicates the relation, such as  VERB for verbal indicators, MOD/HEAD for modifier-head
constructions, etc.), |begin| (character offset for the first character of the predicate 
mention in text), |end| (character offset for the end of the predicate mention in text),
|scale| (if the |Predication| is scalar, the value refers to the |id| of the associated |Scale| element).