Export Output - sequentially

Hi Team, Is there any way I can export out sequentially ?

For Example :
I have 110 records set into Input.jsonl file and at first time I just have annotated first 10 then saved and generated output out of that dataset with command. this will give me output for 10 records only.
python -m prodigy db-out ner_pricing > ./pricing_annotations.jsonl

Next I went into prodigy on same dataset and annotated another 5 records and save and generaged output, this time output is generated for 15 (10 + 5).

My requirement is when I generate output second time I just need to get output for 5 records and NOT for all. Is there any way I can achieve this ?


Hi @ttandel ,

We don't have this capability built-in to db-out. One thing you can do is to separate the JSONL annotations before running prodigy, and save them to different datasets. Then , you call db-out on those datasets to get the split you need.