Refactor build workflows #20
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
|
This change refines the GitHub Actions workflow for handling pull requests by adjusting task names and adding a new testing step for PR review. The workflow now includes both summary and review actions aided by AI, enhancing automation and thoroughness in processing PRs without altering functionality. Walkthrough
Model: gpt-4o | Prompt Tokens: 525 | Completion Tokens: 146 |
|
The CI workflow now references the newer Walkthrough
Model: o3 | Prompt Tokens: 351 | Completion Tokens: 233 | Diff Range: 2a12c4a...458520c |
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
Here’s a friendly code review enriched by AI collaboration. Remember, some suggestions arise from predictive models and might not suit every situation, so select what aligns with your goals. You remain the expert; the AI is merely an ally, ready to support wherever you find it helpful.
Model: o3 | Prompt Tokens: 1141 | Completion Tokens: 1994
.github/workflows/build.yaml
Outdated
There was a problem hiding this comment.
Both testing-summary and testing-review repeat the same boilerplate (runs-on, needs, checkout step, etc.). You can eliminate duplication and improve maintainability by employing a matrix strategy or by using a reusable composite action. Example with a matrix:
pr-jobs:
strategy:
matrix:
include:
- name: PR Summary
uses: ./pr-summary
model: gpt-4o
display: comment-delta
- name: PR Review
uses: ./pr-review
model: gpt-4o
display: review-comment-delta
name: Execute ${{ matrix.name }} action defined in this PR
runs-on: ubuntu-latest
needs: create-release
if: ${{ github.ref_name != 'main' }}
steps:
- uses: actions/checkout@v4
with:
ref: ${{ env.RELEASE_BRANCH }}
- name: Run AI-assisted ${{ matrix.name }}
uses: ${{ matrix.uses }}
with:
aicore-service-key: ${{ secrets.AICORE_SERVICE_KEY }}
model: ${{ matrix.model }}
exclude-files: package-lock.json
display-mode: ${{ matrix.display }}This drastically shortens the workflow file and ensures consistency between both jobs.
There was a problem hiding this comment.
not possible: "In order to enact policies like only using actions defined in the org or repo we can’t allow actions to dynamically change at runtime. So using any sort of dynamic value in uses is not something we will be able to support."
|
Several GitHub Actions workflows now declare explicit Walkthrough
Model: o3 | Prompt Tokens: 989 | Completion Tokens: 293 | Diff Range: 458520c...7187aca |
There was a problem hiding this comment.
Enjoy this AI-supported code review featuring friendly, prediction-based suggestions. Use what resonates with you, adapt or discard the rest, and let your expertise steer the project while AI lends a helpful hand.
Model: o3 | Prompt Tokens: 6168 | Completion Tokens: 4825 | Diff Range: 458520c...7187aca
|
The update shifts our AI-assisted pull-request workflow to the latest major release (v3) of SAP’s GitHub Action and prevents Renovate from repeatedly suggesting version bumps for that action. This simplifies maintenance while keeping CI features intact without affecting runtime code. Walkthrough
Model: o3 | Prompt Tokens: 509 | Completion Tokens: 332 | Diff Range: 7187aca...6cdd925 |
There was a problem hiding this comment.
AI offers a friendly look over your code. It suggests likely improvements, though nothing is set in stone—use only what fits. You remain in charge; AI simply lends a hand.
Model: o3 | Prompt Tokens: 1119 | Completion Tokens: 2151 | Diff Range: 7187aca...6cdd925
…ew actions; change model to o4-mini in build workflow
|
Updated GitHub Actions AI configurations to switch the models used for pull request summaries and reviews. Replaced GPT-4 variants in the ai-assistance workflow with Anthropic Claude 4 Opus, and upgraded from o3 to o4-mini in the build workflow. No functional user-facing changes; this ensures more consistent AI-driven analysis in CI processes. Walkthrough
Model: o4-mini | Prompt Tokens: 514 | Completion Tokens: 910 | Diff Range: 6cdd925...71a8c05 |
There was a problem hiding this comment.
Here’s a supportive code review powered by AI. Think of these insights as suggestions rather than absolute truths—feel free to apply what feels right. You know your project best, so trust your instincts. AI is here to help, but your decisions drive the process.
Model: o4-mini | Prompt Tokens: 1500 | Completion Tokens: 2282 | Diff Range: 6cdd925...71a8c05
.github/workflows/ai-assistance.yaml
Outdated
| model: gpt-4o | ||
| model: anthropic--claude-4-opus |
There was a problem hiding this comment.
The double hyphen in anthropic--claude-4-opus may be a typo. Confirm the model identifier with the action documentation. For clarity, change to a single hyphen if appropriate:
model: anthropic-claude-4-opusThere was a problem hiding this comment.
|
This change updates the GitHub Actions workflow to reference the Walkthrough
Model: o4-mini | Prompt Tokens: 373 | Completion Tokens: 505 | Diff Range: 71a8c05...67367a4 |
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
|
Swapped the AI model in GitHub Actions for PR summary and review workflows to a newer version, aiming to enhance the quality and consistency of automated feedback without altering user-facing features. Walkthrough
Model: o4-mini | Prompt Tokens: 365 | Completion Tokens: 757 | Diff Range: 67367a4...bd13dfb |
There was a problem hiding this comment.
Here's a collaborative code review enhanced by AI assistance. These suggestions offer potential improvements and insights to consider, though they're recommendations rather than requirements. You have full autonomy over your code choices—AI simply provides additional perspective to support your decision-making process. Use whatever feels valuable and aligns with your project goals.
Model: anthropic--claude-4-sonnet | Prompt Tokens: 2759 | Completion Tokens: 1130
| with: | ||
| aicore-service-key: ${{ secrets.AICORE_SERVICE_KEY }} | ||
| model: gpt-4o | ||
| model: anthropic--claude-4-sonnet |
There was a problem hiding this comment.
There appears to be a typo in the model name. 'Claude 4 Sonnet' should likely be 'claude-3-sonnet' based on Anthropic's naming convention:
model: claude-3-sonnet
# or
model: claude-3-5-sonnetThere was a problem hiding this comment.
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
Refactor AI-assisted action execution to use matrix strategy(not possible: "In order to enact policies like only using actions defined in the org or repo we can’t allow actions to dynamically change at runtime. So using any sort of dynamic value in uses is not something we will be able to support.")This change enhances the GitHub Actions workflow security and testing capabilities by adding explicit permissions declarations, upgrading AI model configurations, restructuring test execution, and implementing dependency management controls.
Walkthrough
Model: anthropic--claude-4-sonnet | Prompt Tokens: 1787 | Completion Tokens: 179