EMA NLP

[Firverior02]

This branch includes the full process of natural language processing (NLP).

• A selection of leaflets have been collected
• The leaflets have been retrieved in PDF format
• The Leaflets have been pared and converted to TXT
• All documents provided in both English and Swedish have been used
• The documents have been segmented and validated to make sure SV-EN pairs are formed (not perfect)
• The sentences have been used to further train the Helsinki translation model

[Firverior02]

There are still many bugs, but for the time being it should be good enough