Python project repair - improvement

I need program improvement. The documentation of it is attached. But now program works incorrectly.

Phase 1 - works perfect :)

Phase 2 - not because it uses results of Phase 1 but it should not. It should find simillar articles within one language as it does now but not on results of Phase 1 but on RAW data. The EN portion of data should be downloaded file by file and processed separately because it is not possible to fit it on HDD - just like it is in phase 1. So modifiacation is files it works on.

PHase 3 - now it needs part 2 to finish. But it should be lauched after each EN File processed. Crosslingual sets of articles should be saved in separate files but only the most simillar ones like 5% - not more.

Phase 2 and 3 should be automatic and repeated 99X for each EN FIle.

The code is here: [login to view URL]!AhMECSbJ!hFKXXjib_3Y20VnLlma2L-aTSYa28W1cIsi_ZCdYAsU

Correct EN data is here: http://data.statmt.org/ngrams/docs_en/

C Programming Data Mining Data Processing Perl Python

Project ID: #7898427

About the project