Skip to content

Dev the project

Prepare to run an instance of QGIS-Server-Light

Run it with python and with QGIS installed locally

You can choose this solution if you have QGIS and Python in the correct versions installed on your system.

Preparation

Install a virtual environment for the project like this when pyqgis is installed into your system python like /usr/lib/python3.10/site-packages/ or /usr/lib/python3.10/dist-packages/ depending on your distro.

make install

else you must pass the path to pyqgis (/usr/share/qgis/python) usually:

QGIS_PY_PATH=/usr/share/qgis/python make dev

Run QGIS-Server-Light worker

Spin up a redis instance:

docker run --rm -d -p 1234:6379 --name georama-redis redis

Spin up QGIS-Server-Light worker:

QSL_REDIS_URL=redis://localhost:1234 .venv/bin/python3 -m qgis_server_light.worker.redis

Congratulations... Your QGIS-Server-Light is running and waits for jobs in the redis queue.

Run it within a Docker container

If you don't have QGIS installed or some versions are not matching on you machine the recommended way of running this stack is Docker (docker need to be installed in a recent version).

Preparation

First you need to build the image: 

docker build -t opengisch/qgis-server-light:dev --target dev .

Run QGIS-Server-Light worker

Spin up a redis instance:

docker run --rm -d -p 1234:6379 --name georama-redis redis

Then run: 

docker run --rm -e QSL_REDIS_URL=redis://localhost:1234 -e QSL_LOG_LEVEL=debug -v $(pwd):/app --net host opengisch/qgis-server-light:dev

Congratulations... Your QGIS-Server-Light is running and waits for jobs in the redis queue.

Run it with Docker Compose

Create an .env file and put the content of .env.example into it. Adapt as needed.

docker compose up -d
  • QSL_DATA_MOUNT has to be the valid local absolute path where your qgis projects (and their data) are stored
  • Optional: QSL_LOG_LEVEL log levels as they are defined by pythons logging library
  • Optional: QSL_REPLICAS can be 1 up to * (note, that this has to be balanced with your system resources)
  • Optional: QSL_REDIS_URL=redis://redis (since we are inside our composition, this is optional)

DEV Details

Running things

All deps are transparent. The biggest one is QGIS and its relatives of course. If you want to develop this package further you can use the local way described above.

If you have issues with matching QGIS version on your system and the one used by QGIS-Server-Light you should prefer the docker way with the source code mounted to the running container.

docker run --rm -e QSL_REDIS_URL=redis://localhost:1234 -e QSL_LOG_LEVEL=debug -v $(pwd):/app --net host opengisch/qgis-server-light:dev

That way the current changes you make are transparently available in the container.

For better IDE support it is to be mentioned that the python interpreter inside the container with access to all the necessary python libs is available at opt/qgis-server-light/venv/bin/python3

PyPi Dependencies

If you decide to add new requirements to this project, please consider for which part of the project you need them. Remember, it's split into 3 parts: - exporter (requirements.exporter.txt) - interface (requirements.interface.txt) - worker (requirements.worker.txt)

In general its save to add dependencies to exporter and worker since they are used only in this project. The interface should have as less dependencies as possible and should implement only the minimal understanding of items to be passed around with third party applications through redis. The main goal of all this is to keep the 3rd party application clean of heavy deps.

PyPy Package

The pypi package consists only the part of the interface since it is the part we need to use in 3rd party apps.

PyPi

Not set up yet

GitHub

It is possible to install a package as a dependency in you 3rd party application requirements using a GitHub link.

git+ssh://git@github.com/opengisch/qgis-server-light.git@master#egg=qgis_server_light