[AI] Issue #132: [FEATURE] Structured logs and pipeline health metrics - #133
Conversation
🔍 Automated Code Review✅ SummaryAdded structured logging API with context propagation, step timing, and run summaries. Updated auto-fix script to use new logging functions and handle unhandled rejections. Expanded test coverage for all changes.
|
koydas
added
changes-requested
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 15e89a0fff
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| const systemTokens = estimateTokens(systemPrompt); | ||
| const contextWindow = MODEL_CONTEXT_WINDOW[model] ?? (llmProvider === 'groq' ? 32768 : 200000); |
There was a problem hiding this comment.
The script now terminates right after computing contextWindow, because the entire execution path that builds prompts, calls the LLM, writes generated files, applies attempt labels, and emits outputs was removed. In this state, auto-fix runs become a silent no-op for all PRs that pass earlier guards, so the workflow no longer performs its core function even though it logs startup successfully.
Useful? React with 👍 / 👎.
The PR accidentally deleted the entire core logic after contextWindow computation: token budgeting, diff fetch, LLM call, file writing, and label creation. This restores those 125 lines and adds back the shebang. https://claude.ai/code/session_01ABR3pxJPt1tdUzFLWqmgMq
koydas
added
changes-requested
🤖 Auto-Fix SkippedThis PR modifies |
…gEnd Adds 6 unit tests for the new structured logging functions and 1 integration test verifying the unhandledRejection handler emits run_summary with success: false. Also documents the logging API in code-generation.md. https://claude.ai/code/session_01ABR3pxJPt1tdUzFLWqmgMq
koydas
added
changes-requested
🤖 Auto-Fix SkippedThis PR modifies |
Replaces implicit global crypto.randomUUID() with an explicit named import, compatible with all Node.js ESM environments. https://claude.ai/code/session_01ABR3pxJPt1tdUzFLWqmgMq
koydas
added
changes-requested
🤖 Auto-Fix SkippedThis PR modifies |
2 participants
AI Generated Change
Added log context and summary functionality to logger.mjs and updated auto_fix_pr.mjs to handle unhandledRejection and set log context
Closes #132