PARSEME shared task

Validation procedure of the pilot annotation phase 2 spreadsheets

by Behrang and Federico

9 April 2016

The validation procedure of the spreadsheets which you and your annotators may (or may not) have used during the annotation process does not cover all the annotation problems. Behrang has therefore built a more robust validation script you should try to run for validating all spreadsheets of your language.

Here are the instructions:
1- download the jar file from here (http://typo.uni-konstanz.de/parseme/images/shared-task/validator/ParsemeValidatePhase2.zip) into a directory in your computer and decompress it
2- go to terminal and go to the directory where you have downloaded the jar file (on Windows Open the Command Prompt window by clicking the Start button, clicking All Programs, clicking Accessories, and then clicking Command Prompt.)
3- type the following command: 
java -jar ParsemeValidatePhase2.jar LANGAUGE
(e.g., java -jar ParsemeValidatePhase2.jar English)
4- if you get the sentence "All the annotations seem to be in good order!" you are good. Otherwise you are pointed to the first mistake in one of you language's spreadsheet. Go to that spreadsheet (via the MasterSpreadsheet), fix the problem and repeat the procedure from step 3.

Currently the languages that DO NOT have validation errors are: Polish, German, French, Portuguese, Hungarian.

The languages which still have validation errors are: Italian, Hebrew, Romanian, Swedish, Turkish, English, Maltese, Spanish, Greek.
Please correct the errors ASAP and let me know when you are done. I will then proceed running the script for compiling the MERGED spreadsheet (non editable) and the PLATINIUM spreadsheet (editiable) which will be listed in the MasterSpreadsheet. You should then modify the PLATINIUM spreadsheet by resolving the disagreements together with the other annotators of your language.