Output caching and binary files in DB

This PR contains two major improvements:

1. preparation of binary files for the upload to DB
2. implementation of output caching in the update mode

Caching makes analysis much faster during the wafer scan by reusing output from the previous iterations of the analysis.
More on this in the Wiki

Some minor improvements include:

1. fixed configuration IDs in the last versions, which caused wrong cut_version saved for DB upload
2. added some more tests to the DB output
3. implemented counting of WARNINGs and ERRORs in a configurable way
4. added support of 2FA in db_reader.py
5. added support of development DB and production CROC parts in db_reader.py