This week on the story generation side of things, Vera and I worked on getting the initial state of our stories generated. I finished up necessary infrastructure changes and worked with Andres and Lucas to make sure all of our story data structures have sufficient material to facilitate sentence generation. I started a utilities class for accessing information in the Neo4j database and worked on randomizing and pulling new characters and locations.

Vera worked on action and goal representation in the database and created our “world map”. Our world map is a graph of the locations, where each edge represents a connection between locations and has a compass direction assigned to it.

On the text generation side, Lucas worked on adding adjectives to Locations, Character classes, and Items in the database. Since these are accessed in a similar manner by sentence generation, adjective representation has to be standardized across database items.

Andres worked on creating templates for story introduction. He made separate, interchangeable templates for introducing characters, locations, and goals, and for describing them to the reader. Next week we should be able to finally merge sentence generation and story generation by introducing the complete, initialized world state.

This was also the first week we reported our progress to our adviser Dr. Gonzalez, now that he’s finally back in the office from Europe.

After a productive week, we’re all looking forward to going to Prague this weekend!