🐙 Octopulls, this POC shows how AI acts as a reviewer.

Why Octopulls?

Efficiency in the Review Process

• Automating the generation of code summaries and reviews can save developers and reviewers valuable time, allowing them to focus on more complex issues that require human intervention.

Consistency in Reviews

• AI can provide a level of consistency in code reviews that is difficult to achieve with reviews done exclusively by humans, especially in large teams.

Continuous Learning

• Over time, AI can be trained to identify common error patterns or poor coding practices, making it a learning tool for developers.

Feature	Availability
Code Review	✅
Code Suggestion	✅
Fine-tunning	❌
Commit Suggestion	❌
Code context	❌

Commands

Comando	Descrição
npm install	Install all dependencies
npm run dev	Generate AI reviews of your code integrated with GitHub

Environment variables

Variable	Description	Required
OPENAI_API_KEY	OpenAI API Key (ChatGPT)	✅
GITHUB_AUTH_TOKEN	GitHub Access Token for Applications	✅
GITHUB_ORGANIZATION	Name of the organization or user who will have access to the repository for review	✅
GITHUB_REPOSITORY	Name of the repository that will have the PRs reviewed	✅
GITHUB_PULL_NUMBER	Pull Request number that will have the commits retrieved for review	✅

Product flow

graph TD;
  A[Open PR on GitHub] --> B[GitHub API: Get Latest PR]
  B --> C[GitHub API: Get Commit History]
  C -->|Commits Found| D[Send History to IA]
  C -->|No Commits| E[Feedback: PR with no commits to review]
  D --> F{IA Processes Commits?}
  F -->|Yes| G[IA: Create Commit Summary]
  F -->|Yes| H[IA: Generate Code Review]
  F -->|No| I[Feedback: Failed to process commits]
  G --> J[Generate Comment on PR with Summary]
  H --> J
  I --> K[Add Failed Label on PR]
  E --> L[Add Label with No Commits on PR]

Loading