The workdir in Dockerfile is /opt/namegen

NAMEGEN_FILE ( default: wordlist.csv )
- This file specifies names for generation

NAMEGEN_DB_FILE ( default: namegen.db )
- Database file for namegen
- Hybrid and SQLite drivers only

NAMEGEN_MEMORY_STATE ( default: namegen_memory_state )
- Memory driver state file
- Memory driver only

NAMEGEN_DRIVER ( default: hybrid )
- Options: memory, sqlite
- Defines which driver is user
- Server only

Period of sum of two periodic functions (a,b) is lcm(a,b)

I.E: (wordgen.py)

START.rotate(1) => Period of len(START) = a
END.rotate(1) => Period of len(END) = b

Proof:

When START rotates fully END has rotated a / b laps.
Meaning END will fully rotate when n * a / b is in N.
=> lcm(a,b) * a / b.

I.E: 
Choosing 2,4 we only get 4 distinct possibilities.
BUT: 2,5 => 10 which is much better.

Now this needs to be tweaked as high as possible.

So keep this in mind when tweaking the list.

At the time of writing it was at 
- len(START)=17
- len(END)=16
=> 272

Then numbering uses a LCG with parameters:

- m = 10000
- a = 21
- c = 1
=> period: 10000

Which is used to produce the whole range [0, 9999]

The counts are generated for the whole sum function.
Initially the choice of starting sum is random.
This step uses random.randint(a,b)

The state can be loaded and saved at will.

Since the way LCG's work the state can even be 
recovered from the DB if necessary.

The combined generation limit is set by:
- PERIOD_OF_SUM_FUNCTION * LCG_M

After which the values will repeat.