Makefile

.PHONY: run run-sandbox run-sandbox-container verify-work-run-resume-docker test lint format migrate migrate-new dev-web build-web sync-types dev install db db-stop stop restart k8s-validate k8s-preview-plan k8s-preview-install validate validate-full validate-ci validate-release validate-release-live-backend validate-release-scan verify-api eval-core eval-smoke eval-live eval-agent-scenarios eval-dogfooding-core eval-customer-signal-triage eval-customer-delivery-risk eval-report eval-generate-benchmark eval-live-benchmark eval-sample-workspaces eval-benchmark-samples eval-benchmark-sample-report eval-benchmark-trends eval-benchmark-suite kb-check kb-check-strict kb-check-all kb-snapshot kb-new-answer kb-new-audit

# === Backend ===
run:
	cd server && uv run uvicorn hive.main:app --reload --port 8080

run-sandbox:
	cd server && uv run uvicorn hive.container.sandbox_entrypoint:app --reload --port 8091

run-sandbox-container:
	docker compose --profile sandbox up -d sandbox

verify-work-run-resume-docker:
	-HIVE_SANDBOX_CODING_BACKEND=deterministic_proof docker compose --profile sandbox stop sandbox
	-HIVE_SANDBOX_CODING_BACKEND=deterministic_proof docker compose --profile sandbox rm -f sandbox
	-docker rm -f openhive-sandbox-1
	HIVE_SANDBOX_CODING_BACKEND=deterministic_proof docker compose --profile sandbox up -d --force-recreate sandbox
	server/.venv/bin/python scripts/verify/work_run_resume_docker_live.py

test:
	cd server && uv run pytest tests/ -v --tb=short -m "not integration"

test-e2e:
	cd server && uv run pytest tests/test_e2e/ -v --tb=short -m integration

lint:
	cd server && uv run ruff check hive/ tests/
	cd web && npx eslint src/

format:
	cd server && uv run ruff format hive/ tests/

migrate:
	cd server && uv run alembic upgrade head

migrate-new:
	cd server && uv run alembic revision --autogenerate -m "$(msg)"

# === Database (local dev) ===
db:
	docker compose up -d postgres

db-stop:
	docker compose stop postgres

# === Frontend ===
dev-web:
	cd web && npm run dev

build-web:
	cd web && npm run build

# === Shared types ===
sync-types:
	cd server && .venv/bin/python scripts/generate_dashboard_api_types.py

# === Full stack ===
stop:
	@echo "Stopping backend and frontend..."
	@-pkill -f 'uvicorn hive.main:app' 2>/dev/null || true
	@-pkill -f 'next dev' 2>/dev/null || true
	@sleep 1

restart: stop dev

dev:
	DEBUG_PROMPT=1 $(MAKE) run & $(MAKE) dev-web

install:
	cd server && uv sync --all-extras
	cd web && npm install

validate:
	python3 scripts/validate.py quick

validate-full:
	python3 scripts/validate.py full

validate-ci:
	python3 scripts/validate.py quick --profile ci

validate-release:
	python3 scripts/validate.py full --profile release --scope frontend

validate-release-live-backend:
	python3 scripts/validate.py full --profile release --include-live-backend

validate-release-scan:
	python3 scripts/validate.py full --profile release --scope frontend $(if $(scan_path),--trust-boundary-scan-path "$(scan_path)")

verify-api:
	python3 scripts/verify/api_smoke.py

eval-core:
	python3 scripts/evals/run.py --pack evals/packs/openhive-core-offline.json

eval-smoke:
	python3 scripts/evals/run.py --pack evals/packs/openhive-core-smoke.json

eval-live:
	python3 scripts/evals/run_live_pack.py --pack evals/packs/openhive-live-operator.json --label live

eval-agent-scenarios:
	python3 scripts/evals/run.py --pack evals/packs/openhive-agent-scenarios.json --label scenarios

eval-dogfooding-core:
	python3 scripts/evals/run.py --pack evals/packs/openhive-dogfooding-core.json --label dogfooding-core

eval-customer-signal-triage:
	python3 scripts/evals/run.py --pack evals/packs/openhive-customer-signal-triage.json --label customer-signal-triage

eval-customer-delivery-risk:
	python3 scripts/evals/run.py --pack evals/packs/openhive-customer-delivery-risk.json --label customer-delivery-risk

eval-report:
	@test -n "$(pack)" || (echo "Usage: make eval-report pack=<pack-id>" && exit 1)
	python3 scripts/evals/report.py --pack-id "$(pack)"

eval-generate-benchmark:
	@test -n "$(benchmark)" || (echo "Usage: make eval-generate-benchmark benchmark=<gaia|terminal-bench|swe-bench-verified> source=<path> profile=<path> [pack_id=<id>]" && exit 1)
	@test -n "$(source)" || (echo "Missing required source=<path>" && exit 1)
	@test -n "$(profile)" || (echo "Missing required profile=<path>" && exit 1)
	python3 scripts/evals/generate_benchmark_pack.py --benchmark "$(benchmark)" --source "$(source)" --profile "$(profile)" $(if $(pack_id),--pack-id "$(pack_id)")

eval-live-benchmark:
	@test -n "$(benchmark)" || (echo "Usage: make eval-live-benchmark benchmark=<gaia|terminal-bench|swe-bench-verified> source=<path> project_id=<project-id> [pack_id=<id>]" && exit 1)
	@test -n "$(source)" || (echo "Missing required source=<path>" && exit 1)
	@test -n "$(project_id)" || (echo "Missing required project_id=<project-id>" && exit 1)
	python3 scripts/evals/run_live_benchmark.py --benchmark "$(benchmark)" --source "$(source)" --project-id "$(project_id)" $(if $(pack_id),--pack-id "$(pack_id)") $(if $(profile),--profile "$(profile)") $(if $(workspace_root),--workspace-root "$(workspace_root)") $(if $(workspace_map),--workspace-map "$(workspace_map)") $(if $(room_id),--room-id "$(room_id)") $(if $(room_name),--room-name "$(room_name)")

eval-sample-workspaces:
	python3 scripts/evals/materialize_sample_workspaces.py

eval-benchmark-samples:
	@test -n "$(project_id)" || (echo "Usage: make eval-benchmark-samples project_id=<project-id>" && exit 1)
	python3 scripts/evals/materialize_sample_workspaces.py
	python3 scripts/evals/run_live_benchmark.py --benchmark gaia --source evals/samples/benchmark-sources/gaia-smoke.jsonl --project-id "$(project_id)" --pack-id gaia-smoke-samples --label samples
	python3 scripts/evals/run_live_benchmark.py --benchmark terminal-bench --source evals/samples/benchmark-sources/terminal-bench-smoke.json --project-id "$(project_id)" --workspace-map .artifacts/evals/sample-workspaces/workspace-map.json --pack-id terminal-bench-smoke-samples --label samples
	python3 scripts/evals/run_live_benchmark.py --benchmark swe-bench-verified --source evals/samples/benchmark-sources/swebench-verified-smoke.json --project-id "$(project_id)" --workspace-map .artifacts/evals/sample-workspaces/workspace-map.json --pack-id swebench-verified-smoke-samples --label samples

eval-benchmark-sample-report:
	python3 scripts/evals/report_sample_benchmarks.py

eval-benchmark-trends:
	python3 scripts/evals/report_benchmark_trends.py

eval-benchmark-suite:
	@if [ -z "$(project_id)" ] && [ -z "$(skip_live)" ]; then echo "Usage: make eval-benchmark-suite project_id=<project-id> [include_probes=1] [skip_live=1] [skip_offline=1]"; exit 1; fi
	python3 scripts/evals/run_benchmark_suite.py $(if $(project_id),--project-id "$(project_id)") $(if $(include_probes),--include-probes) $(if $(skip_live),--skip-live) $(if $(skip_offline),--skip-offline) $(if $(label),--label "$(label)") $(if $(api_base_url),--api-base-url "$(api_base_url)") $(if $(room_id),--room-id "$(room_id)") $(if $(room_name),--room-name "$(room_name)")

# === Knowledge base ===
# These KB targets use repo-local wrappers in `scripts/`, while the reusable
# implementation lives in the shared `llm-knowledge-base` Codex skill.
# OpenHive-specific defaults:
# - repo root: this repository
# - KB root resolution order:
#   1. $OPENHIVE_KB_ROOT
#   2. $OPENHIVE_INTERNAL_DOCS_ROOT/kb
#   3. ../openhive-internal-docs/kb
# Prerequisite:
# - optionally install the shared skill under `$CODEX_HOME/skills/llm-knowledge-base`
#   or `~/.codex/skills/llm-knowledge-base`
# - if not installed, these wrappers fall back to `tools/vendor/llm-knowledge-base`
kb-check:
	python3 scripts/kb_status_report.py

kb-check-strict:
	python3 scripts/kb_status_report.py --strict

kb-snapshot:
	python3 scripts/kb_snapshot.py --write

k8s-validate:
	kubectl kustomize deploy/k8s/base >/dev/null
	kubectl kustomize deploy/k8s/jobs/external-postgres-migration >/dev/null
	kubectl kustomize deploy/k8s/overlays/full-gateway-runtime >/dev/null
	kubectl kustomize deploy/k8s/overlays/full-platform-runtime >/dev/null
	kubectl kustomize deploy/k8s/overlays/sandbox-whitelist-example >/dev/null
	sh scripts/k8s/verify-bootstrap.sh

k8s-preview-plan:
	@test -n "$(env_file)" || (echo "Usage: make k8s-preview-plan env_file=deploy/k8s/preview-installer/values.env" && exit 1)
	python3 scripts/k8s/preview_installer.py plan --env-file "$(env_file)"

k8s-preview-install:
	@test -n "$(env_file)" || (echo "Usage: make k8s-preview-install env_file=deploy/k8s/preview-installer/values.env" && exit 1)
	python3 scripts/k8s/preview_installer.py install --env-file "$(env_file)"

kb-check-all:
	python3 scripts/kb_status_report.py && python3 scripts/kb_status_report.py --strict

kb-new-answer:
	@test -n "$(slug)" || (echo "Usage: make kb-new-answer slug=<slug> title=<title> [question=<question>] [scope=<scope>]" && exit 1)
	@test -n "$(title)" || (echo "Missing required title=<title>" && exit 1)
	python3 scripts/kb_create_output.py answer "$(slug)" --title "$(title)" $(if $(question),--question "$(question)") $(if $(scope),--scope "$(scope)")

kb-new-audit:
	@test -n "$(slug)" || (echo "Usage: make kb-new-audit slug=<slug> title=<title> [scope=<scope>]" && exit 1)
	@test -n "$(title)" || (echo "Missing required title=<title>" && exit 1)
	python3 scripts/kb_create_output.py audit "$(slug)" --title "$(title)" $(if $(scope),--scope "$(scope)")