Makefile

PYTHON   := python3
VENV     := .venv
PIP      := $(VENV)/bin/pip
PY       := $(VENV)/bin/python

# ── Default config/output ─────────────────────────────────────────────────────
CONFIG   ?= config.json
BACKEND  ?= faster_whisper
LANGUAGE ?= fr
WORKERS  ?= 2
OUTPUT   ?= ./output
FILE     ?=
URL      ?=

.DEFAULT_GOAL := help

# ── Setup ─────────────────────────────────────────────────────────────────────

.PHONY: install
install: ## Full installation (venv + deps + config)
	bash install.sh

.PHONY: venv
venv: ## Create virtual environment only
	$(PYTHON) -m venv $(VENV)
	$(PIP) install --upgrade pip --quiet

.PHONY: deps
deps: venv ## Install core dependencies
	$(PIP) install --quiet ffmpeg-python tqdm requests yt-dlp pydub

.PHONY: deps-faster-whisper
deps-faster-whisper: deps ## Install faster-whisper backend
	$(PIP) install --quiet faster-whisper

.PHONY: deps-openai
deps-openai: deps ## Install OpenAI backend
	$(PIP) install --quiet openai

.PHONY: deps-all
deps-all: deps deps-faster-whisper deps-openai ## Install all Python dependencies

# ── whisper.cpp ───────────────────────────────────────────────────────────────

WHISPER_CPP_DIR   := whisper.cpp
WHISPER_MODEL     ?= base
WHISPER_MODEL_BIN := $(WHISPER_CPP_DIR)/models/ggml-$(WHISPER_MODEL).bin

.PHONY: whisper-cpp-build
whisper-cpp-build: ## Clone and compile whisper.cpp
	@if [ ! -d "$(WHISPER_CPP_DIR)" ]; then \
	    echo "[+] Cloning whisper.cpp..."; \
	    git clone --depth=1 https://github.com/ggerganov/whisper.cpp $(WHISPER_CPP_DIR); \
	else \
	    echo "[+] whisper.cpp already cloned."; \
	fi
	@echo "[+] Compiling..."
	$(MAKE) -C $(WHISPER_CPP_DIR) -j$$(nproc)
	@echo "[+] Build complete."
	@ls $(WHISPER_CPP_DIR)/main $(WHISPER_CPP_DIR)/build/bin/whisper-cli 2>/dev/null | head -1 | xargs -I{} echo "[+] Binary: {}"

.PHONY: whisper-cpp-model
whisper-cpp-model: ## Download whisper.cpp model (WHISPER_MODEL=base|tiny|small|medium|large-v2)
	@if [ -f "$(WHISPER_MODEL_BIN)" ]; then \
	    echo "[+] Model already exists: $(WHISPER_MODEL_BIN)"; \
	else \
	    echo "[+] Downloading model: $(WHISPER_MODEL)..."; \
	    bash $(WHISPER_CPP_DIR)/models/download-ggml-model.sh $(WHISPER_MODEL); \
	fi

.PHONY: whisper-cpp-setup
whisper-cpp-setup: whisper-cpp-build whisper-cpp-model ## Full whisper.cpp setup (build + model)

# ── Run ───────────────────────────────────────────────────────────────────────

.PHONY: run
run: ## Transcribe FILE= or URL= (e.g. make run FILE=video.mp4)
ifndef FILE
ifndef URL
	$(error Specify FILE=path/to/video.mp4 or URL=https://...)
endif
endif
ifdef FILE
	$(PY) main.py --file "$(FILE)" --backend $(BACKEND) --language $(LANGUAGE) \
	    --workers $(WORKERS) --output-dir $(OUTPUT)
endif
ifdef URL
	$(PY) main.py --url "$(URL)" --backend $(BACKEND) --language $(LANGUAGE) \
	    --workers $(WORKERS) --output-dir $(OUTPUT)
endif

.PHONY: run-config
run-config: ## Run using CONFIG= file (default: config.json)
	$(PY) main.py --config $(CONFIG)

.PHONY: dry-run
dry-run: ## Dry-run — show what would be executed without running
ifdef FILE
	$(PY) main.py --file "$(FILE)" --backend $(BACKEND) --dry-run
else ifdef URL
	$(PY) main.py --url "$(URL)" --backend $(BACKEND) --dry-run
else
	$(error Specify FILE= or URL=)
endif

.PHONY: run-whisper-cpp
run-whisper-cpp: ## Run with whisper_cpp backend (auto-detects binary and model)
	$(eval WCPP_BIN := $(shell ls $(WHISPER_CPP_DIR)/build/bin/whisper-cli $(WHISPER_CPP_DIR)/main 2>/dev/null | head -1))
	@if [ -z "$(WCPP_BIN)" ]; then echo "[x] whisper.cpp binary not found. Run: make whisper-cpp-setup"; exit 1; fi
	@if [ ! -f "$(WHISPER_MODEL_BIN)" ]; then echo "[x] Model not found: $(WHISPER_MODEL_BIN). Run: make whisper-cpp-model"; exit 1; fi
ifdef FILE
	$(PY) main.py --file "$(FILE)" --backend whisper_cpp \
	    --whisper-binary "$(WCPP_BIN)" --whisper-model "$(WHISPER_MODEL_BIN)" \
	    --language $(LANGUAGE) --workers $(WORKERS) --output-dir $(OUTPUT)
else ifdef URL
	$(PY) main.py --url "$(URL)" --backend whisper_cpp \
	    --whisper-binary "$(WCPP_BIN)" --whisper-model "$(WHISPER_MODEL_BIN)" \
	    --language $(LANGUAGE) --workers $(WORKERS) --output-dir $(OUTPUT)
else
	$(error Specify FILE= or URL=)
endif

.PHONY: run-faster-whisper
run-faster-whisper: ## Run with faster_whisper backend
	$(MAKE) run BACKEND=faster_whisper

.PHONY: run-openai
run-openai: ## Run with OpenAI API backend
	$(MAKE) run BACKEND=openai

# ── Output formats ────────────────────────────────────────────────────────────

.PHONY: run-srt
run-srt: ## Transcribe and output SRT subtitles
ifdef FILE
	$(PY) main.py --file "$(FILE)" --backend $(BACKEND) --language $(LANGUAGE) \
	    --format srt --output-dir $(OUTPUT)
else
	$(error Specify FILE=path/to/video.mp4)
endif

.PHONY: run-all-formats
run-all-formats: ## Transcribe and output txt + json + srt + vtt
ifdef FILE
	$(PY) main.py --file "$(FILE)" --backend $(BACKEND) --language $(LANGUAGE) \
	    --format txt,json,srt,vtt --output-dir $(OUTPUT)
else
	$(error Specify FILE=path/to/video.mp4)
endif

# ── Dev ───────────────────────────────────────────────────────────────────────

.PHONY: check
check: ## Validate all imports and CLI
	$(PY) -c "from transcriber.config import TranscriptionConfig; print('config        OK')"
	$(PY) -c "from transcriber.backends.base import TranscriptionBackend; print('backends      OK')"
	$(PY) -c "from transcriber.managers.transcription import TranscriptionManager; print('manager       OK')"
	$(PY) -c "from transcriber.formatters.output import OutputFormatter; print('formatter     OK')"
	$(PY) main.py --help > /dev/null && echo "CLI           OK"

.PHONY: clean
clean: ## Remove output files and temp dirs
	rm -rf output/* __pycache__ transcriber/__pycache__
	find . -type d -name "__pycache__" -exec rm -rf {} + 2>/dev/null || true
	find . -name "*.pyc" -delete

.PHONY: clean-all
clean-all: clean ## Remove venv and output
	rm -rf $(VENV)

# ── Web interface ─────────────────────────────────────────────────────────────

.PHONY: web
web: ## Start web interface at http://localhost:8000
	$(PIP) install --quiet fastapi "uvicorn[standard]" python-multipart
	$(PY) -m uvicorn web.app:app --host 0.0.0.0 --port 8000 --reload

.PHONY: web-install
web-install: ## Install web dependencies only
	$(PIP) install --quiet fastapi "uvicorn[standard]" python-multipart

# ── Help ──────────────────────────────────────────────────────────────────────

.PHONY: help
help: ## Show this help
	@echo ""
	@echo "  Whisper Transcriber"
	@echo ""
	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) \
	    | awk 'BEGIN {FS = ":.*?## "}; {printf "  \033[36m%-22s\033[0m %s\n", $$1, $$2}'
	@echo ""
	@echo "  Variables (override with make VAR=value):"
	@echo "    BACKEND    $(BACKEND)"
	@echo "    LANGUAGE   $(LANGUAGE)"
	@echo "    WORKERS    $(WORKERS)"
	@echo "    OUTPUT     $(OUTPUT)"
	@echo ""
	@echo "  Examples:"
	@echo "    make run FILE=video.mp4"
	@echo "    make run URL=https://youtube.com/... BACKEND=whisper_cpp LANGUAGE=en"
	@echo "    make run-all-formats FILE=video.mp4"
	@echo "    make dry-run URL=https://..."
	@echo "    make check"
	@echo ""