Find file
Fetching contributors…
Cannot retrieve contributors at this time
58 lines (34 sloc) 1.29 KB
Author: Alex Ksikes
Directory structure:
==== data/ ====
The zip files is the actual data from trec
/ flat --- The same data all in the same folder
==== ====
Trec data nornmalized.
Each span is separated by @#@#@ - span_id - @#@#@
spans_id.txt --- Keeps track of each span location
spans_noid.txt --- Used to check if the generated spans match the trec spans
(legalspans.txt should be identical to this file)
==== annotations/ ====
Holds annotations of the corpus
An annotation file has the tab delimited input format:
pmid span_id annotation_id annotation_text
annotation_id is the annotation that will be indexed
annotation_text is the actual text that has been recognized
has an annotation.
==== parsed/ ====
XML/ --- Holds Martinj parsing of the data
sections/sections.clustered --- All sections have extracted and sorted
in order of occurences with count on the left.
categories --- Sections have been assigned to categories.
section_spans --- Offset and length of each section extracted
from the XML.
==== pos/ ====
Holds part of speech tagging from Brown.
==== lucene/ ====
index --- Holds full text index.
index.sections.genes --- Holds full text index + section annotations
+ gene annotations.