connorjoleary / DeepCite

Traversing links to find the deep source of information
GNU General Public License v3.0
69 stars 7 forks source link

Edge cases in tokenizer.py predict #63

Open connorjoleary opened 4 years ago

connorjoleary commented 4 years ago

If the sentence is very small then it may not add legit paragraphs to the predict object. Also, if there are multiple sentences, the one with a link associated would be a better thing to return.

connorjoleary commented 2 years ago

On second thought, the sentence with the link may not be the best to return. What if the sentence with the link just says "Go here." or something like that.