constants.py

LA_COUNTRIES = ["cuba", "dominican republic","puerto rico", "costa rica", "el salvador", "guatemala", "honduras", "mexico", "nicaragua", "panama", "argentina", "bolivia", "chile", "colombia", "ecuador",  "guyana", "paraguay", "peru", "uruguay", "venezuela"]

LA_ETHNICITIES = ["argentinian", "argentine", "cuban", "dominican", "puerto rican", "costa rican", "salvadoran", "salvadorian", "guatemalan", "honduran", "mexican", "nicaraguan", "panamanian", "bolivian", "chilean", "colombian", "ecuadorian", "guyanese", "paraguayan", "peruvian", "uruguayan", "venezuelan", "latino", "latina"] 

NON_LA_ETHNICITIES = ["filipina", "filipino", "italian", "brazilian"]

NOTED_ETHNICITIES = ["spaniard", "castilian", "portuguese", "spanish", "brazilian", "argentinian"]

PLACE_WORDS = ["emigrant", "born", "raised", "from", "grew up"]

DESCENT_WORDS = ["ancestry", "parent", "mother", "father", "descent", "orgin"]

MAIN_TEXT_WORDS =  LA_COUNTRIES + LA_ETHNICITIES + PLACE_WORDS +  DESCENT_WORDS + ["latino", "hispanic"]

CATEGORY_WORDS = LA_ETHNICITIES + LA_COUNTRIES + PLACE_WORDS +  DESCENT_WORDS + ["latino", "hispanic"]