Generate wikipedia page for each ads interest category. Generate keywords for each category with tree structure.
The search_wiki script automatically search through all the ads interests category, and output a json file of all the pageids related to each category.
A database of all terms should be exclude if exist in the wikipedia page title, or the snippets.
Generate lists of keywords for the ads interests category. The script seperate the wikipedia corpse on the server to smaller files, for the sake of preformance. The wikipedia corpse is too large to put there.