kami
d34fdf40aa
The onnxruntime .so was already vendored at deps/onnxruntime-linux-x64-1.26.0 — nothing to download. make eval-router now defaults MAVEN_ONNX_LIB there, so both baselines run by default and only a fresh clone without deps/ falls back to the hash ratchet alone. Prod-representative result, deployed 0.55 gate: 28/76 (36.8%), RU 25/61, EN 3/15, hard 0/11 → 4/11, p50 31ms / p95 71ms. Versus the hash floor's 13/76 at p50 9µs. The finding is not the accuracy, it's the refusal lane: missed clarifies went 0 → 5 of 6. Better embeddings raise cosine everywhere, so the 0.55 threshold that used to hold ambiguous utterances back stops holding — "сделай это" routes to act at 0.847, "бэкап" to chat at 0.755. The gate was implicitly tuned to the hash floor's low similarities. That is an argument about the threshold, not about the embedder, and it lands before #320 rather than after. Also fixes a fixture-model mismatch: ReminderGrammar deliberately skips the extractor at stage 0 and the daemon's applyAction parses the time downstream (stage0.go says so). Charging the router for that slot made 4 exact-match wins read as misses; they are now counted as SlotsDeferred instead. Hash baseline moves 13/76, ratchet to 0.15. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01X5JApcrCRVGmqrxnhynSik
138 lines
5.9 KiB
Makefile
138 lines
5.9 KiB
Makefile
# GOTOOLCHAIN=local pins us to the vendored deps/go tree. Without it, a go.mod
|
|
# `go` directive newer than deps/go re-execs into a downloaded toolchain module,
|
|
# and those ship only 7 of the 15 GOROOT tools (no covdata) -- which makes
|
|
# `test` below fail on the two packages that have no test files. deps-go builds
|
|
# the missing tools in, so the vendored tree is self-sufficient. Keep the version
|
|
# here in step with the `go` directive in go.mod.
|
|
GO_VERSION := 1.25.5
|
|
GO := $(shell pwd)/deps/go/go/bin/go
|
|
export GOTOOLCHAIN := local
|
|
GOFLAGS :=
|
|
CGO_LDFLAGS := -L$(shell pwd)/deps/lib -Wl,-rpath,$(shell pwd)/deps/lib
|
|
CGO_CFLAGS := -I$(shell pwd)/deps/include -I$(shell pwd)/deps/whisper.cpp/ggml/include
|
|
|
|
WHISPER_MODEL := $(shell pwd)/models/stt/ggml-small.bin
|
|
PIPER_BIN := $(shell pwd)/deps/piper/piper
|
|
PIPER_MODEL := $(shell pwd)/models/tts/ru_RU-irina-medium.onnx
|
|
PIPER_ESPEAK := $(shell pwd)/deps/piper/espeak-ng-data
|
|
|
|
.PHONY: all build build-stt build-tts build-daemon build-client build-waked build-web build-poll build-caldav clean test run-stt run-tts run-web download-embedder deps-go eval-router
|
|
|
|
all: build
|
|
|
|
build: build-stt build-tts build-daemon build-client build-waked build-web build-poll build-caldav
|
|
|
|
build-stt:
|
|
CGO_CFLAGS="$(CGO_CFLAGS)" CGO_LDFLAGS="$(CGO_LDFLAGS)" LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
|
|
$(GO) build $(GOFLAGS) -o mavsttd ./cmd/mavsttd/
|
|
|
|
build-tts:
|
|
CGO_CFLAGS="$(CGO_CFLAGS)" CGO_LDFLAGS="$(CGO_LDFLAGS)" LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
|
|
$(GO) build $(GOFLAGS) -o mavttsd ./cmd/mavttsd/
|
|
|
|
build-daemon:
|
|
CGO_CFLAGS="$(CGO_CFLAGS)" CGO_LDFLAGS="$(CGO_LDFLAGS)" LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
|
|
$(GO) build $(GOFLAGS) -o mavend ./cmd/mavend/
|
|
|
|
build-client:
|
|
CGO_CFLAGS="$(CGO_CFLAGS)" CGO_LDFLAGS="$(CGO_LDFLAGS)" LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
|
|
$(GO) build $(GOFLAGS) -o mavenclient ./cmd/mavenclient/
|
|
|
|
build-waked:
|
|
$(GO) build $(GOFLAGS) -o mavwaked ./cmd/mavwaked/
|
|
|
|
build-web:
|
|
$(GO) build $(GOFLAGS) -o mavweb ./cmd/mavweb/
|
|
|
|
build-poll:
|
|
$(GO) build $(GOFLAGS) -o mavpoll ./cmd/mavpoll/
|
|
|
|
build-caldav:
|
|
$(GO) build $(GOFLAGS) -o mavcaldav ./cmd/mavcaldav/
|
|
|
|
run-web: build-web
|
|
./mavweb -addr :9200 -voice 127.0.0.1:9100
|
|
|
|
# Install the vendored Go toolchain from scratch. Go 1.24+ release tarballs no
|
|
# longer ship covdata/pprof/test2json/nm/objdump/trace prebuilt -- `go tool X`
|
|
# builds them on demand, but `go test -coverprofile` calls covdata through
|
|
# base.Tool(), which only stats pkg/tool and exits. So build them in once here.
|
|
GO_TARBALL := go$(GO_VERSION).linux-amd64.tar.gz
|
|
GO_SHA256 := 9e9b755d63b36acf30c12a9a3fc379243714c1c6d3dd72861da637f336ebb35b
|
|
deps-go:
|
|
@mkdir -p deps/go
|
|
cd deps/go && curl -fLO 'https://go.dev/dl/$(GO_TARBALL)'
|
|
cd deps/go && echo '$(GO_SHA256) $(GO_TARBALL)' | sha256sum -c -
|
|
cd deps/go && rm -rf go && tar xzf $(GO_TARBALL) && rm $(GO_TARBALL)
|
|
cd deps/go/go/src && for t in covdata pprof test2json nm objdump trace addr2line buildid; do \
|
|
GOROOT="$(shell pwd)/deps/go/go" $(GO) build -o "$(shell pwd)/deps/go/go/pkg/tool/linux_amd64/$$t" "cmd/$$t" || exit 1; \
|
|
done
|
|
$(GO) version
|
|
|
|
test:
|
|
CGO_CFLAGS="$(CGO_CFLAGS)" CGO_LDFLAGS="$(CGO_LDFLAGS)" LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
|
|
$(GO) test -race -coverprofile=coverage.out ./internal/... ./cmd/...
|
|
|
|
# eval-router — score the held-out RU routing fixture (internal/router/eval).
|
|
# Verbose so the report tables land in the terminal. MAVEN_ONNX_LIB points the
|
|
# prod-representative baseline at the vendored runtime; override it or set it
|
|
# empty to run only the deterministic hash ratchet. This is the measurement
|
|
# Vikunja #319 compares before #320 flips the route decider.
|
|
MAVEN_ONNX_LIB ?= $(shell pwd)/deps/onnxruntime-linux-x64-1.26.0/lib/libonnxruntime.so
|
|
|
|
eval-router:
|
|
MAVEN_ONNX_LIB="$(MAVEN_ONNX_LIB)" $(GO) test -v -count=1 ./internal/router/eval/
|
|
|
|
run-stt: build-stt
|
|
LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
|
|
./mavsttd -socket /tmp/maven/stt.sock -model $(WHISPER_MODEL)
|
|
|
|
run-tts: build-tts
|
|
LD_LIBRARY_PATH="$(shell pwd)/deps/piper" \
|
|
./mavttsd -socket /tmp/maven/tts.sock \
|
|
-piper $(PIPER_BIN) -model $(PIPER_MODEL) -espeak_data $(PIPER_ESPEAK)
|
|
|
|
deps: deps-whisper deps-piper
|
|
|
|
deps-whisper:
|
|
cd deps/whisper.cpp && cmake -B build -DCMAKE_BUILD_TYPE=Release \
|
|
-DWHISPER_BUILD_TESTS=OFF -DWHISPER_BUILD_SERVER=OFF && \
|
|
cmake --build build --config Release -j$$(nproc)
|
|
cp deps/whisper.cpp/build/bin/libwhisper.so* deps/lib/
|
|
cp deps/whisper.cpp/build/bin/libggml*.so* deps/lib/
|
|
cp deps/whisper.cpp/build/bin/libparakeet.so* deps/lib/
|
|
|
|
deps-piper:
|
|
mkdir -p deps/
|
|
curl -sL "https://github.com/rhasspy/piper/releases/download/2023.11.14-2/piper_linux_x86_64.tar.gz" \
|
|
-o /tmp/piper.tar.gz
|
|
tar -xzf /tmp/piper.tar.gz -C deps/
|
|
|
|
EMBEDDER_DIR := $(shell pwd)/models/embedder
|
|
EMBEDDER_MODEL_URL := https://huggingface.co/Xenova/paraphrase-multilingual-MiniLM-L12-v2/resolve/main/onnx/model_quantized.onnx
|
|
EMBEDDER_TOKENIZER_URL := https://huggingface.co/Xenova/paraphrase-multilingual-MiniLM-L12-v2/resolve/main/tokenizer.json
|
|
|
|
download-embedder:
|
|
mkdir -p $(EMBEDDER_DIR)
|
|
curl -sL "$(EMBEDDER_MODEL_URL)" -o "$(EMBEDDER_DIR)/model_quantized.onnx"
|
|
curl -sL "$(EMBEDDER_TOKENIZER_URL)" -o "$(EMBEDDER_DIR)/tokenizer.json"
|
|
@echo ""
|
|
@echo "embedder model downloaded to $(EMBEDDER_DIR)/"
|
|
@echo "To use it, add to mavend.json:"
|
|
@echo ' "voice": {'
|
|
@echo ' ...'
|
|
@echo ' "embedder": {'
|
|
@echo ' "model_path": "$(EMBEDDER_DIR)/model_quantized.onnx",'
|
|
@echo ' "tokenizer_path": "$(EMBEDDER_DIR)/tokenizer.json",'
|
|
@echo ' "lib_path": "/path/to/libonnxruntime.so"'
|
|
@echo ' }'
|
|
@echo ' }'
|
|
@echo ""
|
|
@echo "Install libonnxruntime.so from: https://github.com/microsoft/onnxruntime/releases"
|
|
@echo "e.g. on x86_64 Linux:"
|
|
@echo ' curl -sL "https://github.com/microsoft/onnxruntime/releases/download/v1.15.1/onnxruntime-linux-x64-1.15.1.tgz" | tar xz'
|
|
@echo ' sudo cp onnxruntime-linux-x64-1.15.1/lib/libonnxruntime.so* /usr/local/lib/'
|
|
|
|
clean:
|
|
rm -f mavend mavenclient mavsttd mavttsd mavweb mavpoll mavcaldav mavwaked
|