Update Evaluation

[psaegert]

• [x] Split Evaluation into prediction and computation of the metrics
• [x] Store readable logs of the LLMcoder conversation and analyzers
• [x] Make the evaluation easier to run
• [x] Allow for multiple runs to be executed for a given config
• [x] Improve folder structure of the evaluation results
• [x] Add a variety of interesting config files.

[psaegert]

TODO

• [x] Add unit tests