index attribute for sentences

[jack-rory-staunton]

Feature description

Although technically they are Spans (which do not have indices per se), it would be nice to somehow access the indices of a Doc's sentences:

def get_sentence_index(ent):
    return [i for i, sent in enumerate(doc.sents)
                 if sent.start_char <= ent.start_char 
                 and sent.end_char >= ent.end_char][0]

perhaps by having the generator supply a tuple containing the index.

[svlandeg]

Sentences aren't internally stored as a list, so providing an index in the doc is not straightforward and implementing a function like this would not be efficient.

What you could do though, is loop through all the sentences in list(doc.sents) once, set a custom extension to the corresponding Span, and use that in your other functions.

[jack-rory-staunton]

I'll do that - thanks!