Campaign

Read a campaign config, run the research loop, and read its records. See The campaign config for the config’s sections.

Running a campaign

factor_agent.campaign.Campaign	Run a mining campaign from its config.
factor_agent.campaign.CampaignResult	Every trial record of the run, why it stopped, and where the ledger is.
factor_agent.campaign.Stop	Why the campaign stopped: by is "max_rounds", "max_trials", "patience" or "provider_quota".
factor_agent.preflight.CheckReport	Every Check of a campaign, in order; the run can start when problems is empty.
factor_agent.preflight.Check	One part of a campaign checked: its name, its problems and warnings, or why it was skipped.
factor_agent.cli.main	Run the factor-agent console script; return its exit status.

Configuration

factor_agent.config.CampaignConfig	Every setting of a campaign, read from one YAML file with from_file.
factor_agent.config.DataConfig	The data the Evaluator reads.
factor_agent.config.ScoringConfig	How a factor's raw values become the values that are judged (factor_agent.scoring).
factor_agent.config.NeutralizeConfig	What ranked factors are neutralised against.
factor_agent.config.EvaluatorConfig	How the Evaluator computes the round's graph.
factor_agent.config.GateConfig	The Gate's sandbox limits and graph constraints.
factor_agent.config.AdmissionConfig	What an evaluated candidate needs to be accepted.
factor_agent.config.RoundConfig	The size of a round, the planner's quotas and when the campaign stops.
factor_agent.config.ContextConfig	What the prompt context holds; the run snapshot records it at the first start, and a resume may not change it.
factor_agent.config.AcceptanceConfig	How Acceptance validates, opens the hold-out and exports.
factor_agent.config.RoleConfig	One role's LLM: the backend name, the model and the backend's own settings.
factor_agent.config.Mechanism	A mechanism type a candidate's hypothesis belongs to, with the line the prompts describe it by.

Records and run state

factor_agent.ledger.Ledger	Append and read the trial records of a run directory.
factor_agent.ledger.TrialRecord	One candidate's outcome.
factor_agent.events.Events	The events file of one run directory, appended to by a campaign; safe to use from several threads.
factor_agent.events.read	The events in the file at path, in order; an empty list when there is none.
factor_agent.snapshot.RunSnapshot	The snapshot of a run directory; build it with open or read.
factor_agent.snapshot.SnapshotMismatch	A resume was given files that differ from the run's snapshot.
factor_agent.snapshot.CodeVersionChanged	A resume runs other code than the run was last opened with.
factor_agent.segments.Segments	The training, confirmation, validation and hold-out segments, in that order, and the purge between them.
factor_agent.segments.Segment	A named, inclusive date range.
factor_agent.prompts.Templates	The templates of a campaign: the built-in ones, any of them replaced from a file.
factor_agent.prompts.PromptError	Replaced templates are wrong; the message lists every problem.