RPM processing: file parsing and checksum calculation improvements

Within the new RPM package processing, it is required to parse package file for metadata retirement and calculate the file checksum, both operations required file read and place 2x memory usage.

Need to make changes in current Packages::RPM::ProcessPackageService in order to decrease memory usage.

Related discussions: https://stackoverflow.com/questions/31257577/creating-md5-checksum-while-copying-file-in-rubyhttps://stackoverflow.com/questions/35650220/calculate-hash-from-huge-stream