filter the privacy texts before feeding to datasets

[anjalisoni3655]

• currently, we use nltk library to extract all the texts which fall in the English dictionary.
• But, the sentences may not be correct.
• We need to define our own sentence grammar for that.
• We can also compare each sentence with our existing dataset sentences.