Sentence boundary via detected spans

[justmars]

Attempted component to explicitly prevent improper sentence boundaries via is_sent_end:

@Language.component(name="unsent_spans")
def unsent_spans(doc):
    for span in doc.spans["sc"]:
        for token in span:
            token.is_sent_end = False

Because of error:

ValueError: [E196] Refusing to write to `token.is_sent_end`. Sentence boundaries can only be fixed with `token.is_sent_start`.

Was led to this issue where relevant answer is:

token.is_sent_end is just derived from token.is_sent_start, so you only need to reset is_sent_start to reset all the sentence boundaries.

Am I correct that it's not possible to make token.is_sent_end dependent on span / token attributes? Took a look at Sentencizer but this doesn't seem applicable to the method I'd like to implement. Hoping for suggestions on what path to take. Thanks!