Wasted input evolutions on Golden generation

[DarrenBlackwell]

Am I misinterpreting the concept of evolving the input, overlooking the code, or is this a design flaw?

For a method like synthesizer._a_evolve_input(), it seems that the last evolution is the only one to get used.
I'm not entirely sure why we wouldn't want all the variations of the evolutions. Or why we would run it through the LLM anyway if we aren't going to use it.

Am I missing something obvious re evolving input?

https://github.com/confident-ai/deepeval/blob/main/deepeval/synthesizer/synthesizer.py#L1195

[DarrenBlackwell]

I ended up taking another look at this and it was a complete oversight on my behalf.
The original input gets an evolution applied to it, and subsequent evolutions continue to modify the evolving input, resulting in a single, continuously updated version.