Files
Maven/Dockerfile
T
kami da60c14399 chore: docker, config, delivery sinks, dialogue, and agent docs
- Dockerfile: multi-stage build with CGO_ENABLED=0, embedder model copy,
  non-root user, healthcheck, and /data volume.
- docker-compose.yml: mavend + mavweb services with shared volume, health
  checks, and restart policy.
- .gitignore: ignore models/llm/*.gguf, deploy/telegram.env, tmp artifacts.
- deploy/mavend.json: add LLM, phraser, voice sections (embedder, model
  paths, wake sensitivity). Add telegram token env-var expansion.
- deploy/telegram.env.example: template for telegram bot token.
- internal/config/config.go: add LLM config struct, voice config struct
  (embedder, llama, wake sensitivity), telegram token loading.
- telegramsink: add chat intent delivery support alongside existing types.
- voicesink: skip empty payloads in delivery.
- dialogue/session: add chat intent to anaphora resolution, test coverage.
- AGENTS.md: update with LLM embedder, LFM model download/configure steps,
  new UI conventions.
- REARCH.md: architecture research document.
- cmd/mavend/main.go: wire LLM config, phraser, embedder, telegram config,
  WebAuthn, IPC event/routine handlers, and reactive notes.
2026-07-10 15:49:27 +04:00

110 lines
5.4 KiB
Docker

# syntax=docker/dockerfile:1
#
# Single image, all six daemons. docker-compose runs one container per daemon
# off this image with a different command — the native-lib + toolchain surface
# is shared, so separate images would duplicate ~all of it. Isolation still
# holds: each daemon is its own container/namespace, only mavend mounts the key
# and the db volume.
#
# Native deps are the prebuilt artifacts the repo already carries under deps/
# (libwhisper+ggml-vulkan, onnxruntime, piper/espeak). We do NOT build
# whisper.cpp from source here — COPY the prebuilt .so and headers.
# ponytail: prebuilt-lib copy, not a from-source build. Add a whisper.cpp build
# stage if you ever need reproducibility / a different arch than the host libs.
# trixie, not bookworm: the prebuilt deps/lib/*.so are linked against
# glibc 2.38 + GLIBCXX 3.4.32 (built on the Arch host). bookworm's glibc 2.36
# is too old and the link fails on missing symbols. trixie ships glibc 2.40.
# 1.24 (not 1.23) because golang trixie images start at 1.24; go 1.24 builds the
# `go 1.23` module fine (the directive is a minimum).
FROM golang:1.24-trixie AS build
WORKDIR /src
# libvulkan-dev: libggml-vulkan.so needs libvulkan.so.1 at link time.
RUN apt-get update && apt-get install -y --no-install-recommends libvulkan-dev \
&& rm -rf /var/lib/apt/lists/*
# native build inputs (prebuilt libs + headers), then module cache, then source
COPY deps/lib/ /src/deps/lib/
COPY deps/piper/ /src/deps/piper/
COPY deps/include/ /src/deps/include/
COPY deps/whisper.cpp/ggml/include/ /src/deps/whisper.cpp/ggml/include/
COPY go.mod go.sum ./
RUN go mod download
COPY cmd/ ./cmd/
COPY internal/ ./internal/
# CGO wiring mirrors the Makefile; rpath points at the RUNTIME lib location so
# the binaries find their .so at /opt/maven/lib regardless of LD_LIBRARY_PATH.
ENV CGO_ENABLED=1 \
CGO_CFLAGS="-I/src/deps/include -I/src/deps/whisper.cpp/ggml/include" \
CGO_LDFLAGS="-L/src/deps/lib -L/src/deps/piper -Wl,-rpath,/opt/maven/lib"
RUN go build -o /out/mavend ./cmd/mavend && \
go build -o /out/mavsttd ./cmd/mavsttd && \
go build -o /out/mavttsd ./cmd/mavttsd && \
go build -o /out/mavweb ./cmd/mavweb && \
go build -o /out/mavpoll ./cmd/mavpoll && \
go build -o /out/mavcaldav ./cmd/mavcaldav
# llama.cpp Vulkan build — the phraser/router LFM engine (llama-server). Built
# from source (not a prebuilt vendored blob) so the binary's glibc/GLIBCXX match
# the trixie runtime and GPU offload rides mesa's RADV Vulkan driver — the same
# path whisper already uses on homesrv's AMD iGPU (RADV RENOIR). Pinned to b9601 (parity with
# the host's known-good build). Static (BUILD_SHARED_LIBS=OFF) ⇒ one self-
# contained binary, no libggml/libllama .so to juggle in the runtime; only
# libvulkan.so.1 + libgomp (both already in the runtime) are needed at load.
FROM debian:trixie-slim AS llama
ARG LLAMA_REF=b9601
RUN apt-get update && apt-get install -y --no-install-recommends \
ca-certificates git cmake build-essential libvulkan-dev \
glslc glslang-tools spirv-headers spirv-tools \
&& rm -rf /var/lib/apt/lists/*
RUN git clone --depth 1 --branch ${LLAMA_REF} \
https://github.com/ggml-org/llama.cpp /src/llama.cpp
WORKDIR /src/llama.cpp
RUN cmake -B build \
-DCMAKE_BUILD_TYPE=Release \
-DGGML_VULKAN=ON \
-DBUILD_SHARED_LIBS=OFF \
-DLLAMA_CURL=OFF \
-DLLAMA_BUILD_SERVER=ON \
&& cmake --build build --config Release -j"$(nproc)" --target llama-server
FROM debian:trixie-slim AS runtime
# tzdata so the TZ env (set in compose) resolves — otherwise Go can't load the
# zone and time.Now() stays UTC, and mavend answers clock/date queries and
# evaluates quiet-hours in UTC.
RUN apt-get update && apt-get install -y --no-install-recommends \
ca-certificates libvulkan1 mesa-vulkan-drivers libgomp1 tzdata \
python3 python3-pip && \
pip3 install --no-cache-dir --break-system-packages 'dateparser==1.4.1' && \
apt-get purge -y --auto-remove python3-pip && \
rm -rf /var/lib/apt/lists/*
# runtime native libs: whisper/ggml (incl. vulkan) are real files in deps/lib.
COPY deps/lib/ /opt/maven/lib/
COPY deps/piper/ /opt/maven/piper/
# The router embedder needs onnxruntime 1.26, but deps/lib only carries dangling
# symlinks to it (they point at an absolute HOST path that isn't in the image).
# Ship the real 1.26 lib and repoint the symlinks. Piper keeps its OWN
# onnxruntime 1.14 under /opt/maven/piper (found via piper's $ORIGIN runpath +
# its exact soname libonnxruntime.so.1.14.1) — do NOT copy piper's .so into
# /opt/maven/lib, that clobbers 1.26 and mavend dies with "ORT API base: 2".
COPY deps/onnxruntime-linux-x64-1.26.0/lib/libonnxruntime.so.1.26.0 /opt/maven/lib/
RUN cd /opt/maven/lib \
&& ln -sf libonnxruntime.so.1.26.0 libonnxruntime.so \
&& ln -sf libonnxruntime.so.1.26.0 libonnxruntime.so.1
COPY --from=build /out/ /opt/maven/bin/
# the LFM engine: static Vulkan llama-server on PATH; the phraser spawns it by
# name (bin_path "llama-server"). GPU offload needs /dev/dri passed to mavend.
COPY --from=llama /src/llama.cpp/build/bin/llama-server /opt/maven/bin/
ENV LD_LIBRARY_PATH=/opt/maven/lib PATH=/opt/maven/bin:$PATH
# unprivileged; core owns the key + db, modules own nothing.
RUN useradd -r -u 10001 -m maven \
&& mkdir -p /run/maven /var/lib/maven \
&& chown maven:maven /run/maven /var/lib/maven
USER maven
WORKDIR /opt/maven