Repository navigation
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Depends on #521 (stacked branch). The CLI-only change is commit 3f6cb1c; the preceding env-import commit is reviewed in #521. Please merge that dependency first; this branch will need rebasing onto main after a squash merge.
Partially addresses Direction A of #468. Connect the existing experiment-dictionary and dotenv converters to the local example runner with mutually exclusive
--alphaevolve-config(JSON) /--alphaevolve-envoptions. Require an explicit OE endpoint and, for env files, an instructions file. Optional--api-key-envselects a key by environment-variable name.Imported configurations default to seed evaluation;
--evolveexplicitly starts model calls. Use the imported generation budget unless a positive--iterationsreduces it. Preserve legacy OE YAML--configevolution and its one-iteration default. Validate option/import errors before loading example code, and document the modes and budget limitations. Core loaders, upstream seeds/evaluators and cloud entrypoints are unchanged.Validation:
python -m unittest discover -s tests -q: 727 tests passed, including 12 new runner tests, subprocess CLI invocations and an offline controller/worker evolution.git diff --checkpass.674dd5e558e917320c251c085e47fae33e1bbcff. Seed-only modes made zero model requests. Each example then completed one worker iteration through the real OpenAI-compatible client against a local HTTP fixture with a fixed proposal (env import for circle packing, JSON for the other two). Original source hashes were unchanged. External connections were blocked; no paid model or Google Cloud calls were made. This validates interoperability, not learned optimization quality.No SDK emulation, exporter, concurrency equivalence or hard cost/request cap is claimed.
AI disclosure: Codex assisted with implementation, tests and documentation for the user's selected follow-up to #468. All reported validation was executed locally.