fix: Incorrect added token can cause issues when adding token as multiword token

[stephantul]

In the current version of model2vec it is possible to attempt adding a multiword added token if the token is already present in the vocabulary. This happens in the following circumstances:

1. The tokenizer has a token that is incorrect: a token which is pre-tokenized into multiple subwords.
2. We attempt to add precisely that token as a multiword unit.

The logic bug is as follows: we don't check whether multiword tokens are already present in the original tokenizers vocabulary. Therefore, we try to add the same token twice. This assumption is sound, because normally a tokenizer's vocabulary can't contain tokens that are preprocessed into multiple tokens.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
model2vec/distill/distillation.py	90.00% <100.00%> (-0.15%)	⬇️
model2vec/tokenizer/tokenizer.py	100.00% <100.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.