jedarden/ai-code-battle
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
51 commits 1 branch 0 tags 358 MiB
Commit graph

3 commits

Author SHA1 Message Date
jedarden
76e8791e4d Add evaluation arena, promotion gate, and retirement policy (Phase 7) 
- arena/arena.go: 10-match mini-tournament running candidate as a local
  subprocess against diverse live opponents sampled across the rating
  distribution; AES-GCM secret decryption for opponent auth
- arena/psro.go: Nash equilibrium computation for the 1×K meta-game;
  FictitiousPlayNash included for future K×K support
- arena/winrate.go: Wilson-score 95% CI for win-rate calculation; draws
  counted as 0.5 wins
- arena/gate.go: two-part promotion gate — Nash value ≥ threshold AND
  MAP-Elites niche fill or improvement; detailed reason strings
- promoter/promoter.go: full promotion pipeline — bot source + Dockerfile
  + K8s Secret/Deployment/Service manifests, docker build, git commit/push
  (ArgoCD sync), kubectl readiness poll, bots-table INSERT, programs-table
  update; RetireBot and EnforcePolicy (rating threshold + population cap 50)
- db/db.go: add bot_name / bot_secret migration columns
- db/programs.go: ListPromoted, SetBotNameAndSecret, UnsetPromoted,
  GetByBotID, PromotedCount helpers for promotion/retirement lifecycle
- main.go: evaluate and retire subcommands wiring arena + gate + promoter;
  remove unused island flag from evaluate
- arena/arena_test.go: 21 unit tests covering Nash, Wilson CI, Gate logic,
  and selectDiverse opponent sampling
- promoter/promoter_test.go: tests for Dockerfiles, bot-ID/secret generation,
  AES-GCM helpers, and K8s manifest templates

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-03-26 23:32:37 -04:00
jedarden
5669688984 Add validation pipeline, sandbox, and evolution DB layer (Phase 7) 
Three-stage fail-fast validator for LLM-generated bot candidates:
- syntax.go: language-aware parse (go/parser for Go; py_compile, rustfmt,
  tsc, javac, php -l for others; brace-balance fallback)
- schema.go: regex detection of /health + /turn endpoints and "moves" field
- sandbox.go: nsjail-isolated smoke test — builds bot, polls /health, sends
  5 signed /turn requests, verifies JSON moves responses
- validator.go: orchestrates stages with fail-fast short-circuit

DB layer:
- programs table + CRUD (create, get, list, updateFitness, setPromoted)
- validation_log table with RecordValidation, IslandPassRates,
  IslandValidationStats for per-island pass-rate tracking
- seed.go: 6 generation-0 bots across alpha/beta/gamma/delta islands

MAP-Elites grid (mapelites/grid.go): 2-D behavior grid on aggression×economy
axes; TryPlace keeps the fittest occupant per niche.

acb-evolver CLI gains two new subcommands:
  validate <file> -lang <lang> [-island <island>] [-nsjail] [-nolog]
  validation-stats (tabular per-island pass-rate breakdown)

cmd/acb-api/db.go: add programs table to API schema so the API can query
promoted evolved bots.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-03-26 22:45:13 -04:00
jedarden
bd4b0d3244 Add LLM prompt builder and ensemble integration (Phase 7) 
- selector: tournament selection for parent sampling from island populations
- prompt: assembles evolution prompts from parent code, replay analysis, and meta description
- llm: OpenAI-compatible client routing to ZAI proxy with fast (GLM-5-Turbo) and strong (GLM-5) tiers, plus code block extraction from model responses
- Tests for prompt assembly, code extraction, and tournament selection

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-03-26 22:26:09 -04:00