fermilink.optimize.campaign module

Campaign entrypoints for optimize mode.

fermilink.optimize.campaign.read_campaign_status(args)

Parameters:

args (Namespace)

Return type:

dict[str, Any]

fermilink.optimize.campaign.run_campaign(args)

Parameters:

args (Namespace)

Return type:

dict[str, Any]

fermilink.optimize.campaign.run_goal_campaign(args)

Goal mode: parse goal.md, analyse source, generate benchmark, then run campaign.

This is the entry point for fermilink optimize goal.md when the markdown has goal-structured sections (or --goal is passed).

Steps: 1. Parse goal.md into a structured spec. 2. Run source-analysis agent turn → structured JSON analysis. 3. Run benchmark-generation agent turn → benchmark.yaml + runner.py. 4. Validate generated files. 5. Optionally retry generation if validation fails. 6. Fall through to run_campaign() with generated benchmark.

Parameters:

args (Namespace)

Return type:

dict[str, Any]

fermilink.optimize.campaign.run_quick_campaign(args)

Parameters:

args (Namespace)

Return type:

dict[str, Any]