The payload contains only the bounding boxes information and the link to the temporary location of the pdf used for the labeling.
The entities objects are missing.
And I cannot see an easy solution to extract the labels using the entity characters index of the text entities when I will have them if I have only the pdf file.
The platform should provide the labeled entities using the classical SPAN format, with at minimum the following information:
- text
- category
- start
- end
By the way, the documentation page about the bounding boxes payload format seems deprecated when I compare with the json file I get with the API call.
Thanks
Jerome