Creating topics over time fails if model has deleted topics

Nobody has claimed this yet.

### Have you searched existing issues? 🔎

- [x] I have searched and found no existing issues

### Desribe the bug

If the model has some of its topics deleted (via delete_topics), calling topics_over_time fails with error:
```
ValueError Traceback (most recent call last)
Cell In[37], line 1
----> 1 topics_over_time = topic_model.topics_over_time(tweets, timestamps, nr_bins=20)

File ~/miniconda3/envs/tgram/lib/python3.13/site-packages/bertopic/_bertopic.py:900, in BERTopic.topics_over_time(self, docs, timestamps, topics, nr_bins, datetime_format, evolution_tuning, global_tuning)
896 selection = documents.loc[documents.Timestamps == timestamp, :]
897 documents_per_topic = selection.groupby(["Topic"], as_index=False).agg(
898 {"Document": " ".join, "Timestamps": "count"}
899 )
--> 900 c_tf_idf, words = self._c_tf_idf(documents_per_topic, fit=False)
902 if global_tuning or evolution_tuning:
903 c_tf_idf = normalize(c_tf_idf, axis=1, norm="l1", copy=False)

File ~/miniconda3/envs/tgram/lib/python3.13/site-packages/bertopic/_bertopic.py:4451, in BERTopic._c_tf_idf(self, documents_per_topic, fit, partial_fit)
4448 if fit:
4449 self.ctfidf_model = self.ctfidf_model.fit(X, multiplier=multiplier)
-> 4451 c_tf_idf = self.ctfidf_model.transform(X)
4453 return c_tf_idf, words

File ~/miniconda3/envs/tgram/lib/python3.13/site-packages/sklearn/utils/_set_output.py:319, in _wrap_method_output..wrapped(self, X, *args, **kwargs)
317 @wraps(f)
318 def wrapped(self, X, *args, **kwargs):
--> 319 data_to_wrap = f(self, X, *args, **kwargs)
320 if isinstance(data_to_wrap, tuple):
321 # only wrap the first output for cross decomposition
322 return_tuple = (
323 _wrap_data_with_container(method, data_to_wrap[0], X, self),
324 *data_to_wrap[1:],
325 )

File ~/miniconda3/envs/tgram/lib/python3.13/site-packages/bertopic/vectorizers/_ctfidf.py:113, in ClassTfidfTransformer.transform(self, X)
110 if self.reduce_frequent_words:
111 X.data = np.sqrt(X.data)
--> 113 X = X * self._idf_diag
115 return X

File ~/miniconda3/envs/tgram/lib/python3.13/site-packages/scipy/sparse/_matrix.py:55, in spmatrix.__mul__(self, other)
54 def __mul__(self, other):
---> 55 return self._matmul_dispatch(other)

File ~/miniconda3/envs/tgram/lib/python3.13/site-packages/scipy/sparse/_base.py:911, in _spbase._matmul_dispatch(self, other)
909 if issparse(other):
910 if N != other.shape[0]:
--> 911 raise ValueError(
912 f"{err_prefix} (n,k={N}),(k={other.shape[0]},m)->(n,m)"
913 )
914 return self._matmul_sparse(other)
916 # If it's a list or whatever, treat it like an array

ValueError: matmul: dimension mismatch with signature (n,k=30856),(k=30855,m)->(n,m)
```

### Reproduction

```python
from bertopic import BERTopic
import re
import pandas as pd

# Prepare data
trump = pd.read_csv('https://drive.google.com/uc?export=download&id=1xRKHaP-QwACMydlDnyFPEaFdtskJuBa6')
trump.text = trump.apply(lambda row: re.sub(r"http\S+", "", row.text).lower(), 1)
trump.text = trump.apply(lambda row: " ".join(filter(lambda x:x[0]!="@", row.text.split())), 1)
trump.text = trump.apply(lambda row: " ".join(re.sub("[^a-zA-Z]+", " ", row.text).split()), 1)
trump = trump.loc[(trump.isRetweet == "f") & (trump.text != ""), :]
timestamps = trump.date.to_list()
tweets = trump.text.to_list()

topic_model = BERTopic(verbose=True)
topics, probs = topic_model.fit_transform(tweets)

topic_model.delete_topics([7])

topics_over_time = topic_model.topics_over_time(tweets, timestamps, nr_bins=20) # throws the error
topic_model.visualize_topics_over_time(topics_over_time, top_n_topics=20)

```

### BERTopic Version

0.17.4

Start by reproducing the supplied example with BERTopic 0.17.4, then inspect the topics_over_time entry point in bertopic/_bertopic.py and the transform path in bertopic/vectorizers/_ctfidf.py. Trace the topic and c-TF-IDF dimensions after delete_topics([7]); done means topics_over_time completes and the subsequent visualization call works without the dimension-mismatch error.

Written by the indexing model from the issue text.