Text Filtering Tool Based on Article Needed

Here is some nice artivle I found: [login to view URL] The idea is to filter from text corpora sentences that are irrelevan in some text subject domain. For example it is not necessary to have sport comments in medical texts. I would need an implelentation of such tool. For LM training I would like to use [login to view URL] tool. In their solution they filter only monoligual data. I would also like the program to filter bi-lingual. For example if I want to filter [login to view URL] file based on [login to view URL] I remove poor lines from somedata.en. But also if I want to filter [login to view URL] and [login to view URL] files based on [login to view URL] I remove poor lines from [login to view URL] and corresponding line in somedata.fr. We can assume that one line represents one sentence in such files.