Converting prodigy jsonl to spacy: am I missing something?

[Zatteliet]

Supposing I have a set of documents annotated with Prodigy, so they are in the output format as in examples in the docs. I want to convert these to Spacy Docs.

So far I've been doing this manually, with e.g.

def jsonl2doc(jsonl_dict: dict, nlp: Language) -> Doc:
    doc = nlp(jsonl_dict["text"])
    ents = []
    for ann in jsonl_dict.get("spans", []):
        span = Span(
            doc,
            ann["token_start"],
            ann["token_end"] + 1,
            ann["label"],
        )
        ents.append(span)
    doc.set_ents(ents)
    return doc

It works, but it feels like either prodigy or spacy is missing some built-in functionality around this, given how tightly the libraries are couples otherwise. Is there for obvious way of solving this task that I'm missing?

[thomashacker]

Hey,
If you annotated the data with prodigy, you can use the data-to-spacy command (https://prodi.gy/docs/recipes#data-to-spacy) to get the .spacy files for training.