The correct way to represent this would have not been by focus, but by decision.
Give Turkey a decision that becomes available once Germany reaches a surrender progress roughly corresponding to their historical situation in February 1945, somewhere around 80-90%. Another requirement would be for Turkey to not already be in a war.
Once AI Turkey clicks this decision, activate an AI strategy that prevents it from sending forces outside Turkish territory and home waters for the remainder of the war. To prevent exploiting the AI, this strategy is abandoned if another power declares war on Turkey whilst it is active, although preferably you'd still have an AI strategy active that makes AI TUR send forces to fight against its aggressor, rather than against Germany/other Axis members.
The reward of the decision would be something like political power, enough to make the temporary stability hit from war worth it, or perhaps stability as well, and better relations with certain Allied countries. Other effects would obviously be that Turkey joins the Allies and the war.
Or if this is not the desired approach, have the same decision as above, rewards and all, but abstract the formal declaration of war out completely and keep Turkey at peace in-game, replacing the state of war with a serious relations hit with Germany and an embargo to prevent trade. That might even be the best way to model it.
Either approach would have far more preferable to what the developers ended up doing.
EDIT: You could even have this remain a focus with the same requirements and effects, but I do think a decision would have been preferable.