# Minimal helpers for the Gemma 4 stack.  Run `make help`.
ENV     ?= dev-cpu
ARGS    ?=
COMPOSE  = docker compose

.PHONY: help env models pull up up-gpu down ps logs test clean

help: ## Show this help
	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | \
	  awk 'BEGIN{FS=":.*?## "}{printf "  \033[36m%-10s\033[0m %s\n",$$1,$$2}'

env: ## Copy profiles/$(ENV).env -> .env   (ENV=dev-cpu|cpu-server|cpu-server-xl|gpu)
	cp profiles/$(ENV).env .env

models: ## Download required BF16 mmproj (use ARGS=--weights to also fetch GGUFs)
	./scripts/download-models.sh $(ARGS)

pull: ## Pre-pull docker images
	$(COMPOSE) pull

up: ## Start the CPU stack
	$(COMPOSE) up -d

up-gpu: ## Start the GPU stack (needs profiles/gpu.env)
	$(COMPOSE) -f docker-compose.yml -f docker-compose.gpu.yml up -d

down: ## Stop the stack
	$(COMPOSE) down

ps: ## Show service status
	$(COMPOSE) ps

logs: ## Tail logs (SVC=llama-swap to filter)
	$(COMPOSE) logs -f --tail=120 $(SVC)

test: ## Run the e2e smoke tests
	./scripts/e2e-test.sh

clean: ## Stop and remove containers + volumes
	$(COMPOSE) down -v