why SpanGroup?

[afriedman412]

Is there any reason to use a SpanGroup instead of just a list of Spans if it isn't explicitly required by the code?

[jkgenser]

The purpose is that you may need to disambiguate separate groups. It's a way of keeping groups of spans organized. E.g. there may be one spacy component that creates spangroups under key foo and another component that creates spangroups under key bar. SpanGroup allows you to distinguish.

[adrianeboyd]

The main motivation for SpanGroup was to have a data structure containing spans that can be saved efficiently as part of the Doc, with Doc.to_bytes or DocBin or pickle.

It's extremely inefficient to save independent Span objects, since they'd each need a full doc+vocab to have all the information they'd need. spacy actually prevents you from pickling spans because it's not a good idea, with this error:

File "spacy/tokens/span.pyx", line 206, in spacy.tokens.span.Span.__reduce__
NotImplementedError: [E112] Pickling a span is not supported, because spans are only views of the parent Doc and can't exist on their own. A pickled span would always have to include its Doc and Vocab, which has practically no advantage over pickling the parent Doc directly. So instead of pickling the span, pickle the Doc it belongs to or use Span.as_doc to convert the span to a standalone Doc object.

In a script where you're just temporarily working with spans and you don't need to save them, then a list of spans is fine.

[afriedman412]

awesome thanks!!!