Add Swahili Language Support for Sentence Parsing and Text Processing

[m453h]

Media Cloud does not need any special language support to collect data for a given language. However, the data collected will not undergo any further processing to provide more insights e.g. Listing top word counts on the sources application for collections which predominantly contain Swahili language will mostly show stop words.

Thus, we would like to add support for the Swahili language in the system which will enable features like tokenization, stemming and stop words removal for Swahili which will greatly improve the data returned by various word-counting API endpoints.