Our system has been tested in the following environment:

OS: Linux (Ubuntu 14.04)
Memory: 64GB RAM
CPU: Intel Xeon
Prerequisite software/library: Python2.7

We release our system for cross-lingual text stream alignment.
To use our system, you should first extract the content from the compressed file using the following command:

tar -zxvf [filename]

Then, you'll get the folder called crosskn_codes. In this folder, you should preprare cross-lingual coordinated streams you want to align in the folder ch_stream and en_stream. The depth of the data folders should be 3 (i.e., the data folder -> subfolders (e.g., months) -> files). Due to upload limitation and the copyright issue, we only keep 4 sample documents in ch_stream and en_stream folder.

Now, you can run the script to align the stream using the following command:

bash script.sh

After all the commands finish, you'll get a file called pro4_fullscore_decode.txt.maxscore. This file is the final alignment results which rank the alignments by their overall score defined in Section 3.5 in the paper submission.
