I am somehow facing a strange problem using regex for urls mentioned in crawl-urlfilter.txt. Before using any regx for urls, I test them in a standalone class and they work correctly i.e. pattern.matcher(url).find() returns true.
But when the same url and regex is used during crawling, it returns false. I am not sure how it behaves differently.
Let me give an example
My link DB is growing substantially now and Im' crawling some 12 million
urls a day. I plan on generating my linkdb in portions (10) of 100 million
each to place on my sand box servers for a distributed search cluster.
Before I move this out of hadoop and place it on local file systems I want
to filter my linkdb for any adult content.
Does anyone have any pointers or ready made filters for this?
I'm sure I can create some filters to do this to a degree; however a tried
and true system would be ideal.