One quick solution could indeed be a regular expression as suggested in the comments if you can afford to look-up all your data and see what easy rule would be sufficient.
If you have a lot of variety in your data, take advantage of a proxy task: sentence tokenization. In fact, if you manage to split sentences, you're basically done.
For that, don't reinvent the wheel, use available sentence tokenizers:
>>> from nltk.tokenize.punkt import PunktSentenceTokenizer
>>> tokenizer = PunktSentenceTokenizer()
>>> sentences = tokenizer.tokenize("The temperature today is 30.8 degrees celsius. However yesterday at 12:00 A.M., M. John said it was 27.1 degrees.")
>>> print(sentences)
['The temperature today is 30.8 degrees celsius.',
'However yesterday at 12:00 A.M., M. John said it was 27.1 degrees.']
Getting rid of full stops becomes very easy: just remove the final dot if there's one:
>>> print([s[:-1] for s in sentences if s.endswith(".") else s])
['The temperature today is 30.8 degrees celsius',
'However yesterday at 12:00 A.M., M. John said it was 27.1 degrees']
Hope that helps.