This README.txt gives details of the supplementary material for 
ACL 2018 paper submission ID: 1337 
(titled: "Identification of Alias Links among Participants in Narratives").


Contents of the SupplementaryMaterial directory (dir) are as follows:
======================================================================
1. README.txt (this file)
2. Submission_1337_Appendix.pdf
3. "Data" dir
4. "Scripts" dir

Details of Data:
================
The "Data/gold" directory contains gold-standard annotations for the four history datasets.
Filenames are in the following format:  "Participants_<datasetID>_gold.conll"

The "Data/prediction" directory contains the predictions of the baselines B1, B2 as well as our method (M) in the CoNLL format for the four history datasets. 
Filenames are in the following format:  "Participants_<datasetID>_<approachName>.conll"

"datasetID" is one of {BoH, Fas, Mao, Nap}. 

Details of "datasetID" naming convention as well as the baselines B1, B2 and our method (M) are discussed in the paper.


Details of "Scripts":
=====================

Following steps need to be run in the given order to transform the annotations of ACE 2005 dataset to the required format.

Step 1:
=======
1.a) Ensure that following jar files are present in the same dir as that of ACE_DataReader.java.
	slf4j-api.jar, 
	slf4j-simple.jar, 
	stanford-corenlp-3.7.0.jar, 
	stanford-corenlp-3.7.0-models.jar

1.b) Compile the ACE_DataReader.java using following command:
	javac -cp "*" ACE_DataReader.java 

1.c) Run the following command to create intermediate files in "output" dir (note: the empty "output" dir must exist already). 

	java -cp "*:." ACE_DataReader <Path_to_ACE_2005_Parent_dir>/ace_2005_td_v7/data/English/nw/timex2norm   output/


Step 2:
=======
Run the following the command to convert the intermediate files created in step 1.c to CoNLL format. (note: the empty  "CoNLL_Dir" dir must exist already). 
	
	python convert_ACE_to_CoNLL_format.py outputTest/  CoNLL_Dir/ fileOrder.txt

At the end of Step 2, ACE2005.gold.conll would get created in CoNLL_Dir.
It contains the desired transformed ACE2005 annotations in CoNLL format.


