Moving Diplomacy SpeechLines to data files

[BottledByte]

Moving Diplomacy SpeechLines to data files

[BottledByte]

@tuomount I see you have already started offloading SpeechLines to JSON. You seem to use my old design I posted in #666, where file is treated as object.
I have to stop you there. Don't do that, for reasons mentioned above. I have tried it already and it is a bad direction.

I don't have much time lately, but I will look into the SpeechSet implementation when possible.
After all, I have functional one already 😄. It just has some design flaws and cannot handle the "ID problem".

But I would appreciate if you could fix that dichotomy between "behavior" and "text ID" in SpeechFactory.
Adding SpeechType like "ACKNOWLEDGE_WAR" might be sufficient. 👍
(and then slightly refactoring SpeechFactory)

Then, I will be able to put SpeechSets together much faster and in much better form.
I don't want to convert data back and forth, when I can prevent the situation by "doing it well from the start". 😐

[tuomount]

Reason why I did the JSON first, I wanted to try it if it could work out, but then I noticed reading speeches in that format would not be easy that DataLoader. It would need to have extra class something like SpeechSet.

I can add add the ACKNOWLEDGE_WAR SpeechType.

[BottledByte]

but then I noticed reading speeches in that format would not be easy that DataLoader

Exactly, DataLoader is designed to work with objects. My design from #666 predates DataLoader... and it was done with jackson-jr 😄 . Plus there are issues with further extensibility of SpeechSet not-yet-defined class.

I can add add the ACKNOWLEDGE_WAR SpeechType.

👍
It is not possible to do this without it anyway. Or some other solution to the ID issue. 🙂

[BottledByte]

I have checked-out the changes, but the issue with "one ID for two texts" is still not fixed. TBH, I have no idea why the changes with ACKNOWLEDGE_WARE were done 😄, but having 2 IDs point to one text is completely OK. Might be used for flavor texts even.

However, the true issue is SpeechFactory.createAgreeWithWarLine(SpaceRace) method, as it's call is not paired with any "ID", but rather called ad-hoc in code. Also, it returns "SpeechType.AGREE". If it will be addressed by "ID" (SpeechType), the factory should return that SpeechType and code should accept it.

One more issue, in the same spirit of "I request A, I get B" is with Mechions and Synthdroids, which "alter" output of SpeechFactory.createDeclineAngerLine() from SpeechType.DECLINE_ANGER to SpeechType.DECLINE. This is unexpected behavior, and it violates several assumptions about SpeechFactory. And, which is the most problematic part, it causes SpeechFactory define behavior. This is another blocker to dehardcoding SpeechLines and should be remedied as well.

[tuomount]

Okay, I did not first get what you meant by ACKNOWLEDGE_WAR. But yes, now createAgreeWithWarLine is called as any other speechline and has it's own type, which is ACKNOWLEDGE_WAR.

Removed also special thing that robots could not get angry.

[BottledByte]

Seems to work, closing.