Implement Character Repetition Correction [resolves #268]

[ertugrul-dmr]

• Added optional mention, hashtag, emoji like character repetition normalizer for tokenizers and text2doc.
• Normalizer works like how @onatyap proposed in issue #268
• It takes strings like "Süppeeeer", "Berbaaat", "Muhteşeemmmm" and returns "Süper", "Berbat", "Muhteşem" in tokenization part.
• Added test cases.
• I tested it on several prebuilt models and here are the results:

Prebuilt Model	Original Result	Preprocessed Result
Tweet Sentiment Classification	3-Fold F-1: 0.8640, 5-Fold F-1: 0.8669	3-Fold F-1: 0.8587 5-Fold F-1: 0.8640
Movie Review Sentiment Classification	F-1: 0.8258	F-1: 0.8242
Telco Tweet Sentiment Classification	F-1: 0.6871, Accuracy: 0.6925	F-1: 0.696, Accuracy: 0.691
Turkish Customer Reviews Classification	F-1: 0.851	F-1: 0.852

• This method might be useful in some cases where data comes in high number of repeated character words.
• Apart from test results @onatyap could you check the code and confirm that what you proposed in the issue?