arcgis-python-api icon indicating copy to clipboard operation
arcgis-python-api copied to clipboard

JSON Format Mismatch for NER data export from Doccano Editor vs. ArcGIS Format

Open Daremitsu1 opened this issue 1 year ago • 0 comments

Describe the bug The JSON format exported from Doccano's text editor for labeling tasks is currently structured with keys "id", "text", and "label", whereas ArcGIS expects the format to contain "labels".

Might not be a bug per se, and by changing it to labels it is working. But since Doccano is being advertised as the primary text editor for ArcGIS annotations if this could be adjusted, would be great.

To Reproduce Steps to reproduce the behavior:

  1. Create annotations in doccano and export them.
  2. Read the data using prepare_data from arcgis.learn.

Expected behavior ArcGIS should be able to parse JSON files exported from Doccano without any modifications, allowing users to seamlessly import labeled data for further analysis and visualization.

Platform (please complete the following information):

  • OS: [e.g. iOS]
  • Browser [e.g. chrome, safari] chrome
  • Python API Version [e.g. 1.6.2] (you can get this by typing print(arcgis.__version__) 1.8.4

Additional context Add any other context about the problem here, attachments etc.

Daremitsu1 avatar Feb 26 '24 18:02 Daremitsu1