Chinedum Echeta
Chinedum Echeta
This PR should close #157, by using the commit SHA and ref from the pull request when using `workflow_run`, instead of from the default branch.
Closes #159 Use `label` instead of `ref` when fetching the base and head for comparing commits. This includes the repo owner, e.g. `MishaKav:main` instead of just the branch name, so...
When using `workflow_run` or `workflow_dispatch`, the base and head are determined directly from the branches of the PR. However, this does not work for forks, because the head branch is...
`workflow_run` runs in the context of the default branch, therefore when constructing URLs, it is incorrectly using the commit SHA from the default branch instead of from the pull request....
Closes #758 ## Purpose * This PR adds Semantic Kernel as a third orchestration choice * Semantic Kernel is being used as a wrapper around OpenAI functions, therefore users should...
Closes #790 ## Purpose * Updates default OpenAI deployment to `gpt-35-turbo-16k`, which supports more tokens for the bigger prompt. ## Does this introduce a breaking change? ``` [ ] Yes...
# Describe the bug A clear and concise description of what the bug is. Following #648 , the size of the prompt has increased massively, and exceeds the number of...
Required by #320 # Description Add a new orchestration strategy, `semantic_kernel`. It should use function calling to call the existing functions. # Tasks _To be filled in by the engineer...
Required by #320 # Description Replace the existing functions, which use LangChain, to use Semantic Kernel functions that do not rely on LangChain. # Tasks _To be filled in by...