neat/species/core/augmentation

Concrete extended-history augmentation helpers for the species core chapter.

The root species.core.ts file decides whether augmentation should run at all. This companion file owns the narrower mechanics that actually walk the history rows, find matching live species records, and copy derived structural summaries into entries that are still missing them.

The implementation deliberately stays conservative:

• it enriches only rows that lack the extended fields,
• it skips history rows whose species no longer exist in the live registry,
• it derives only compact summary values instead of copying full genomes or connection lists into history.

neat/species/core/augmentation/species.core.augmentation.ts

backfillExtendedHistoryEntries

backfillExtendedHistoryEntries(
  history: SpeciesHistoryEntry[],
  context: SpeciesHistoryBackfillContext,
): void

Backfill missing extended history fields in place.

This chapter owns the concrete augmentation workflow so the root species.core.ts file can stay focused on policy decisions such as whether the backfill should run at all.

The helper walks recorded generations first and species rows second. For each row, it performs a small three-step decision:

1. skip rows that already contain the extended metrics,
2. look up the matching live species by id,
3. if that species still exists and has members, summarize its innovation coverage and copy the result into the history row.

This makes the backfill useful for read-time teaching and diagnostics without pretending that old history can always be reconstructed perfectly from the current runtime state.

Parameters:

• history - - Recorded species history to enrich.
• context - - NEAT context exposing current species and optional fallback innovations.

Returns: Nothing. The history entries are mutated in place when backfill succeeds.

SpeciesHistoryBackfillContext

Runtime surface required to backfill extended species-history fields.

The augmentation chapter only needs the current live species registry plus an optional innovation fallback for legacy connection records.

That narrowness is intentional. The backfill path is a read-side enrichment step, not a second controller facade, so it should depend only on the live species evidence and the small fallback hook needed to interpret older connection records.