hjess/mmd
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
b38702e082c3fdaa6307776aae92979aeb009fa4
mmd/Dockerfile

38 lines
804 B
Docker
Raw Normal View History

perf: cache pip downloads and Docker layers in CI build - Add BuildKit pip cache mount to Dockerfile (persists torch/transformers downloads across builds on the runner host) - Enable DOCKER_BUILDKIT=1 in workflow env - Pull previous image before build for layer cache reuse - Add --cache-from latest + BUILDKIT_INLINE_CACHE=1 to docker build BUILDKIT_INLINE_CACHE embeds cache metadata in the pushed image so --cache-from works from the registry (cold runner or new runner). SHA tag push reuses already-pushed layers - only metadata transferred. Expected improvement: - Unchanged code: ~5 min -> ~30 sec (layer cache hit) - requirements changed: ~5 min -> ~1 min (pip disk cache) - Push: only changed layers uploaded; SHA tag is free
2026-05-27 22:21:15 +02:00
# syntax=docker/dockerfile:1
feat: containerize for mmd.i80.dk deployment via Gitea/Nomad - Add Dockerfile (python:3.12-slim, HF_HOME=/app/data/hf-cache) - Add mmd.nomad (multi-task: web=dashboard, worker=scheduler) - Add .gitea/workflows/deploy.yml (build->Harbor, deploy->Nomad) - Add scheduler.py (stdlib scheduler replaces cron in container) - Add requirements.txt - dashboard.py: LOG_DIR + PORT/HOST from env vars - saxo_auth.py: TOKEN_FILE from SAXO_TOKEN_FILE env var - .gitignore: proper ignores for container project Volume moneymaker-data (/app/data) holds: - logs/ (shared between web+worker) - .saxo_token.json (pre-copy once after first deploy) - hf-cache/ (HuggingFace FinBERT cache) Gitea secrets required: DATABASE_URL, ANTHROPIC_API_KEY, SAXO_APP_KEY, SAXO_APP_SECRET_1, HARBOR_ROBOT_TOKEN
2026-05-26 22:30:38 +02:00
FROM python:3.12-slim
ARG BUILD_VERSION=unknown
ARG BUILD_TIME=unknown
ARG GIT_COMMIT=unknown
ENV PYTHONDONTWRITEBYTECODE=1 \
PYTHONUNBUFFERED=1 \
BUILD_VERSION=${BUILD_VERSION} \
BUILD_TIME=${BUILD_TIME} \
GIT_COMMIT=${GIT_COMMIT} \
HF_HOME=/app/data/hf-cache \
TRANSFORMERS_VERBOSITY=error \
HF_HUB_DISABLE_PROGRESS_BARS=1
WORKDIR /app
perf: cache pip downloads and Docker layers in CI build - Add BuildKit pip cache mount to Dockerfile (persists torch/transformers downloads across builds on the runner host) - Enable DOCKER_BUILDKIT=1 in workflow env - Pull previous image before build for layer cache reuse - Add --cache-from latest + BUILDKIT_INLINE_CACHE=1 to docker build BUILDKIT_INLINE_CACHE embeds cache metadata in the pushed image so --cache-from works from the registry (cold runner or new runner). SHA tag push reuses already-pushed layers - only metadata transferred. Expected improvement: - Unchanged code: ~5 min -> ~30 sec (layer cache hit) - requirements changed: ~5 min -> ~1 min (pip disk cache) - Push: only changed layers uploaded; SHA tag is free
2026-05-27 22:21:15 +02:00
RUN --mount=type=cache,target=/root/.cache/pip \
pip install --upgrade pip
feat: containerize for mmd.i80.dk deployment via Gitea/Nomad - Add Dockerfile (python:3.12-slim, HF_HOME=/app/data/hf-cache) - Add mmd.nomad (multi-task: web=dashboard, worker=scheduler) - Add .gitea/workflows/deploy.yml (build->Harbor, deploy->Nomad) - Add scheduler.py (stdlib scheduler replaces cron in container) - Add requirements.txt - dashboard.py: LOG_DIR + PORT/HOST from env vars - saxo_auth.py: TOKEN_FILE from SAXO_TOKEN_FILE env var - .gitignore: proper ignores for container project Volume moneymaker-data (/app/data) holds: - logs/ (shared between web+worker) - .saxo_token.json (pre-copy once after first deploy) - hf-cache/ (HuggingFace FinBERT cache) Gitea secrets required: DATABASE_URL, ANTHROPIC_API_KEY, SAXO_APP_KEY, SAXO_APP_SECRET_1, HARBOR_ROBOT_TOKEN
2026-05-26 22:30:38 +02:00
COPY requirements.txt ./
perf: cache pip downloads and Docker layers in CI build - Add BuildKit pip cache mount to Dockerfile (persists torch/transformers downloads across builds on the runner host) - Enable DOCKER_BUILDKIT=1 in workflow env - Pull previous image before build for layer cache reuse - Add --cache-from latest + BUILDKIT_INLINE_CACHE=1 to docker build BUILDKIT_INLINE_CACHE embeds cache metadata in the pushed image so --cache-from works from the registry (cold runner or new runner). SHA tag push reuses already-pushed layers - only metadata transferred. Expected improvement: - Unchanged code: ~5 min -> ~30 sec (layer cache hit) - requirements changed: ~5 min -> ~1 min (pip disk cache) - Push: only changed layers uploaded; SHA tag is free
2026-05-27 22:21:15 +02:00
RUN --mount=type=cache,target=/root/.cache/pip \
pip install -r requirements.txt
feat: containerize for mmd.i80.dk deployment via Gitea/Nomad - Add Dockerfile (python:3.12-slim, HF_HOME=/app/data/hf-cache) - Add mmd.nomad (multi-task: web=dashboard, worker=scheduler) - Add .gitea/workflows/deploy.yml (build->Harbor, deploy->Nomad) - Add scheduler.py (stdlib scheduler replaces cron in container) - Add requirements.txt - dashboard.py: LOG_DIR + PORT/HOST from env vars - saxo_auth.py: TOKEN_FILE from SAXO_TOKEN_FILE env var - .gitignore: proper ignores for container project Volume moneymaker-data (/app/data) holds: - logs/ (shared between web+worker) - .saxo_token.json (pre-copy once after first deploy) - hf-cache/ (HuggingFace FinBERT cache) Gitea secrets required: DATABASE_URL, ANTHROPIC_API_KEY, SAXO_APP_KEY, SAXO_APP_SECRET_1, HARBOR_ROBOT_TOKEN
2026-05-26 22:30:38 +02:00
COPY . .
RUN mkdir -p /app/data/logs /app/data/hf-cache
ENV PORT=5001 \
HOST=0.0.0.0 \
LOG_DIR=/app/data/logs \
SAXO_TOKEN_FILE=/app/data/.saxo_token.json
EXPOSE 5001
CMD ["python", "dashboard.py"]
Reference in New Issue Copy Permalink