Most of these parsers are really bad! Because they insert data, one row at a time. You are better off, parsing the data into tab separated files and use the database (LOAD DATE INFILE), instead of killing you server for 12 to 24 hours, you can load all the data in 2 or 3 hours. Doing it this way will also automatically sort the index(s), if you are going to do fulltext searching on all the data, instead of just searching titles! This will save you another 1 or 2 hours of process time!
yj!