def find(self): print("Finding similarity") simi = Similarity() embeddings_index, knn = simi.similarity() print("***************")
from similarity import Similarity a = Similarity(update=False) print(a.similarity('A caza de papil é linda', 'A Casa de papel é linda ')) print(a.similarity('A caza de papel é linda', 'A Casa de papel é linda ')) print(a.similarity('A casa de papel eh linda', 'A Casa de papel é linda '))
print( f"*** Loaded {sim.num_words} word vectors with dimensionality {sim.num_features}" ) print("** Computing matches") n = 0 n_unknown = 0 for reviewer in Author.objects.filter( Q(volunteer=True) | Q(first_author=True)): if Bid.objects.filter(author=reviewer).exists(): continue email = reviewer.email try: reftexts = reference[email] except KeyError: print( f"##### No reference text found for {reviewer.email} ({reviewer.first_name} {reviewer.last_name})" ) n_unknown += 1 continue for paper in Paper.objects.all(): text = "\n\n".join([paper.title, paper.abstract]) s = sim.similarity(text, "\n\n".join(reftexts)) Bid.objects.create(paper=paper, author=reviewer, score=0, weight=s) n += 1 print( f"*** Assigned similarity scores for {n} reviewers (skipped {n_unknown} reviewers)" ) #print(reviewer, reviewer.email in reference)