You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
IkarisGPT currently relies on repository description for injecting a context/understanding of the project in the LLM prompt. Repository descriptions are often brief and does not provide a verbose understanding of the project. For IkarisGPT to produce better descriptions, it should be capable of encapsulating useful information about the project from different sources (like readme.md) or even provide for the maintainer to manually enter important details of the project in a summary.
Tasks
Research for improved ways of injecting context into prompts.
Implement a mechanism to inject facts from README.md to prompt.
(Optional) Implement a mechanism for the maintainers to manually input project specific details.
Analyse and present the difference this makes in description generation.
Expected Outcome
Improved description generation with more project specific details, less tangential and more issue specific texts in overall context of the project.
The text was updated successfully, but these errors were encountered:
Description
IkarisGPT currently relies on repository description for injecting a context/understanding of the project in the LLM prompt. Repository descriptions are often brief and does not provide a verbose understanding of the project. For IkarisGPT to produce better descriptions, it should be capable of encapsulating useful information about the project from different sources (like readme.md) or even provide for the maintainer to manually enter important details of the project in a summary.
Tasks
Expected Outcome
Improved description generation with more project specific details, less tangential and more issue specific texts in overall context of the project.
The text was updated successfully, but these errors were encountered: