scietex.vacuum_gauge_service

scietex.vacuum_gauge_service is a scietex.service worker that monitors one or more Modbus/TCP vacuum gauges through the Scietex Modbus gateway. Each gauge is polled through a scietex.hal.vacuum_gauge driver selected by dotted path, and the per-gauge readings are published to Valkey. The service accepts monitoring commands and per-gauge control commands through the framework's task-handler channel.

The worker is a long-running daemon. It seeds its in-memory gauge list from vacuum_gauge.yml at startup and updates it from the framework's remote-config channel at runtime.

Python ≥ 3.10 · License: MIT

Configuration

The service reads vacuum_gauge.yml from a vacuum_gauge/ subdirectory of its config directory. When the file is missing, the service generates a default on first run:

gateway:
  host: 127.0.0.1
  port: 502
  timeout: 1.0
gauges: []

Each gauge entry:

Running

start-vacuum-gauge-service --conf-dir /etc/scietex

--conf-dir defaults to the framework's auto-resolved config directory when omitted. The service runs in the foreground until it receives SIGINT or SIGTERM.

Environment variables:

Container image

The Containerfile is a two-stage build on python:3.14-slim-trixie:

• Builder stage installs the released scietex.vacuum_gauge_service wheel from PyPI (selected by the VERSION build arg) and its dependencies into /opt/venv, then slims the venv (drops pip/setuptools/wheel, bytecode caches, and the unused per-interpreter glide_shared extensions).
• Runtime stage copies the venv, creates a non-root appuser (uid 1000), prepares /config, and runs start-vacuum-gauge-service.

The image installs the published package rather than building from the local checkout, so the image and the PyPI artifact for a given version are the same code. The VERSION build arg must name a version that exists on PyPI.

The service reaches gauges over Modbus/TCP through a gateway, so the container needs no serial devices and no privileged ports.

The image sets these defaults:

ENV SCIETEX_CONFIG_DIR=/config \
    SCIETEX_SERVICE_NAME=VacuumGaugeService \
    SCIETEX_LOGGING_LEVEL=INFO
VOLUME ["/config"]

Build

podman build --build-arg VERSION=0.1.1 -t scietex-vacuum-gauge-service .

build_image.sh builds a multi-arch manifest (linux/amd64,linux/arm64) and pushes it to registry.buro-nts.ru/scietex-vacuum-gauge-service, tagged with the version from version.py. It passes that version as the VERSION build arg, so the image always installs the matching PyPI release. Pass --latest to also push the latest tag.

Run

podman run --rm \
  -v scietex-config:/config \
  -e SCIETEX_SERVICE_NAME=VacuumGaugeService \
  scietex-vacuum-gauge-service:0.1.1

The gateway address and gauge list come from the API domain plugin, which publishes them to Valkey; vacuum_gauge.yml under the config volume only seeds the initial values.

Development

Dependencies are managed with uv; checks and tests run through tox:

uv sync --all-extras
uv run tox