Installation¶
The installation is using the Python distribution system Miniconda (default installation of no conda
found)
to maintain software dependencies. Any conda
installation should work the same. To use a pre-installed conda
distribution, simply make sure that it can be found on the shell path.
Requirements¶
The installation works on Linux 64 bit distributions (tested on Ubuntu 16.04).
Note
Windows is not officially supported, but some patches have been applied to help using it. If you find some problems, please submit a new issue or open a pull request with fixes.
From GitHub Sources¶
Install weaver as normal user from GitHub sources:
git clone https://github.com/crim-ca/weaver.git
cd weaver
make install
If no conda
environment is activated, the install
process will setup a new or reuse the conda environment
named weaver
and install all dependency packages. If an environment is activated, Weaver will be installed in
that environment. You can also enforce a specific environment using:
make CONDA_ENV=<my-env> install
Windows Installation¶
Minimal support is provided to run the code on Windows. To do so, the Makefile
assumes you are running in a
MINGW
environment, that conda
is already installed, and that it is available from CONDA_HOME
variable or
similar. If this is not the case, you will have to adjust the reference variables accordingly.
Note
Windows support is not official and any dependency could stop supporting it at any given time. Particularly, libraries for Celery task execution have a tendency to break between versions for Windows. The application is regularly evaluated on a Linux virtual machine. It is recommended to run it as so or using the existing Docker images.
Known issues¶
Package
shapely.geos
has C++ dependency togeos
library. If the package was installed in aconda
environment, but throughpip install
call, the source path will not be found. You have to make sure to install it usingconda install -c conda-forge shapely
.The example weaver.ini.example file uses
gunicorn
by default to take advantage of its performance features, but this package does not support Windows. Alternatively, you might need to usewaitress
by replacing it in the[server:main]
section.
Please refer to Configuration and Running sections for following steps.