Add more complex integration test cases to ensure the functionality of the Finalizer. Making the tests green will most likely involve improving the prompt of the Finalizer.
Background
The Finalizer was moved into the Supervisor subgraph in the issue #170. The basic functionality of the Finalizer was covered in integration tests. However, some of the functionality suggested in the prompt could not be seen in more complex tests cases. To rectify this:
Considerations/Recommendations for Integration tests:
[ ] Add a syntax scoring evaluation function. It should give a bad score if:
[ ] Any agent names are in the final response
[ ] Extra headers/sub-headers are added to the final response
Test case(s):
[ ] Two agents, one responding with invalid answers, and the other responding with valid answers.
[ ] The valid answers should be included in the final response.
[ ] The invalid answers should be removed from the final response.
[ ] Apology for the invalid answers should be included in the final response.
Acceptance Criteria(s):
[ ] Confirm with the PO/Team the desired behaviour
Current behaviour: If the answer is wrong then the Finalizer apologies. But if the answer is partially correct then in the final answer it only includes the partially correct part and skips the incorrect part.
[ ] Come up with the necessary integration test cases
[ ] Make the tests green by improving the Finalizer prompt
Description
Add more complex integration test cases to ensure the functionality of the Finalizer. Making the tests green will most likely involve improving the prompt of the Finalizer.
Background
The Finalizer was moved into the Supervisor subgraph in the issue #170. The basic functionality of the Finalizer was covered in integration tests. However, some of the functionality suggested in the prompt could not be seen in more complex tests cases. To rectify this:
Considerations/Recommendations for Integration tests:
Test case(s):
Acceptance Criteria(s):