Script to parse WebAnno XMI data and find differences in annotations
Uses the python packages csv and xml. If you dont already have them installed, get them via pip.
The script takes to xmi-export files from Webanno and puts out a log file in the .csv format. Invoke the script via:
$ python3 compare_annotations.py <file_1>.xmi <file_2>.xmi <log>.csv
The script will save a log file under the specified name. This file ist structured in the following way (with an example line included):
| Sentence | Frame | Key | Value File 1 | Value File 2 |
|---|---|---|---|---|
| 2 | geht es | Label | Communication | Feeling |
Due to technical reasons, when comparing Frame Links, a 1-to-1-correspondence between the files cannot always be found. In this case, the script will log a list of possible links in the other file to leave it up to the user to decide for the match.