Skip to content

Export to CoNLL 2012 format has duplicate (V*) tags #1432

@reckart

Description

@reckart

Original issue: webanno/webanno#1535


Describe the bug
When tagging sentences for semantic role labeling using the SemPred and SemArg layers of WebAnno, the export to CoNLL 2012 format duplicates the SemPred as (V*) to every column of the predicate arguments columns. The associated arguments (ARG0, ARG1 etc.) are not duplicated and present in the correct column.

This also happens when importing a correctly labeled file in CoNLL 2012 format a exporting it right after.

Exported file (I aligned the columns to provide better readability):

#begin document (test.txt); part 000
test.txt   0   0        Die            -            -         -    -   -          -         *          *                      *  -
test.txt   0   1 Studierenden          -            -         -    -   -          -         *          *                      *  -
test.txt   0   2     kennen            -            -         -  null   -          -         *        (V*)                   (V*) -
test.txt   0   3  Verfahren            -            -         -    -   -          -         *     (ARG1*                 (ARG1*  -
test.txt   0   4        der            -            -         -    -   -          -         *          *                      *  -
test.txt   0   5    Analyse            -            -         -    -   -          -         *          *                      *  -
test.txt   0   6          ,            -            -         -    -   -          -         *          *                      *  -
test.txt   0   7 Konzeption            -            -         -    -   -          -         *          *                      *  -
test.txt   0   8        und            -            -         -    -   -          -         *          *                      *  -
test.txt   0   9 Verifikation          -            -         -    -   -          -         *          *                      *  -
test.txt   0  10        von            -            -         -    -   -          -         *          *                      *  -
test.txt   0  11 Geschäftsmodellen     -            -         -    -   -          -         *          *)                     *) -
test.txt   0  12        und            -            -         -    -   -          -         *          *                      *  -
test.txt   0  13     können            -            -         -    -   -          -         *          *                      *  -
test.txt   0  14      diese            -            -         -    -   -          -         *          *                      *  -
test.txt   0  15        auf            -            -         -    -   -          -         *          *                 (ARG0*  -
test.txt   0  16   gegebene            -            -         -    -   -          -         *          *                      *  -
test.txt   0  17 Fallbeispiele         -            -         -    -   -          -         *          *                      *  -
test.txt   0  18        und            -            -         -    -   -          -         *          *                      *  -
test.txt   0  19     eigene            -            -         -    -   -          -         *          *                      *  -
test.txt   0  20 Projektideen          -            -         -    -   -          -         *          *                      *) -
test.txt   0  21   anwenden            -            -         -  null   -          -         *        (V*)                   (V*) -
test.txt   0  22          .            -            -         -    -   -          -         *          *                      *  -

#end document

To Reproduce
Steps to reproduce the behavior:

  1. Annotate a sentence with SemPred and define associated SemArgs
  2. Click on the export icon
  3. Choose format "CoNLL 2012"
  4. Click export
  5. Exported file has duplicate (V*) tags over all predicate argument columns

Expected behavior
The SemPred layers represented as (V*) should only be applied to their corresponding columns, such as in this file:

#begin document (test.txt); part 000
test.txt   0   0        Die            -            -         -    -   -          -         *          *                      *  -
test.txt   0   1 Studierenden          -            -         -    -   -          -         *          *                      *  -
test.txt   0   2     kennen            -            -         -  null   -          -         *        (V*)                    *  -
test.txt   0   3  Verfahren            -            -         -    -   -          -         *     (ARG1*                 (ARG1*  -
test.txt   0   4        der            -            -         -    -   -          -         *          *                      *  -
test.txt   0   5    Analyse            -            -         -    -   -          -         *          *                      *  -
test.txt   0   6          ,            -            -         -    -   -          -         *          *                      *  -
test.txt   0   7 Konzeption            -            -         -    -   -          -         *          *                      *  -
test.txt   0   8        und            -            -         -    -   -          -         *          *                      *  -
test.txt   0   9 Verifikation          -            -         -    -   -          -         *          *                      *  -
test.txt   0  10        von            -            -         -    -   -          -         *          *                      *  -
test.txt   0  11 Geschäftsmodellen     -            -         -    -   -          -         *          *)                     *) -
test.txt   0  12        und            -            -         -    -   -          -         *          *                      *  -
test.txt   0  13     können            -            -         -    -   -          -         *          *                      *  -
test.txt   0  14      diese            -            -         -    -   -          -         *          *                      *  -
test.txt   0  15        auf            -            -         -    -   -          -         *          *                 (ARG0*  -
test.txt   0  16   gegebene            -            -         -    -   -          -         *          *                      *  -
test.txt   0  17 Fallbeispiele         -            -         -    -   -          -         *          *                      *  -
test.txt   0  18        und            -            -         -    -   -          -         *          *                      *  -
test.txt   0  19     eigene            -            -         -    -   -          -         *          *                      *  -
test.txt   0  20 Projektideen          -            -         -    -   -          -         *          *                      *) -
test.txt   0  21   anwenden            -            -         -  null   -          -         *         *                    (V*) -
test.txt   0  22          .            -            -         -    -   -          -         *          *                      *  -

#end document

Screenshots
image

Please complete the following information:

  • Version and build ID: build from source, commit 381cc0f939b0747f6ded33ba534ebe6e065c16fe
  • OS: Arch Linux
  • Browser: Chrome 78.0.3904.108

Additional context
none

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

Milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions