flooding

Introduction

Usage, etc.

More details in src/flooding/USAGE.txt .

Works with RabbitMQ version 2.8.7.

Install production / staging server

Linux task machine (i.e. task 200). The task server checks out the master trunk of flooding.

Init

$ bin/fab staging_taskserver init

Update

$ bin/fab staging_taskserver update_task

Init

$ bin/fab production_taskserver init

Update

$ bin/fab production_taskserver update_task

See staging-task-200.cfg as an example, it actually serves tasks 200, 210 and 220.

Problems can arise when installing netcdf4. Try:

$ sudo apt-get install libhdf5-serial-dev libnetcdf-dev

WARNING: buildout run on jupiter

Buildout chokes if you run it with the smb share still mounted. So, as root, first unmount ('umount') the share:

#> umount /srv/flooding.lizardsystem.nl/var/external_data

After buildout finished correctly, mount it again as root:

#> mount /srv/flooding.lizardsystem.nl/var/external_data

Development installation

The first time, you'll have to run the "bootstrap" script to set up setuptools and buildout:

$> python bootstrap.py

And then run buildout to set everything up:

$> bin/buildout

(On windows it is called bin\buildout.exe).

You'll have to re-run buildout when you or someone else made a change in setup.py or buildout.cfg.

The current package is installed as a "development package", so changes in .py files are automatically available (just like with python setup.py develop).

If you want to use trunk checkouts of other packages (instead of released versions), add them as an "svn external" in the local_checkouts/ directory and add them to the develop = list in buildout.cfg.

Tests can always be run with bin/test or bin\test.exe.

Workflows

The next workflow_templates are created on migration:

DEFAULT_TEMPLATE_CODE = 1 (workflow for a scenario with sobek model) IMPORTED_TEMPLATE_CODE = 2 (workflow for a scenario with onknown model via import) THREEDI_TEMPLATE_CODE = 3 (workflow for scenario with 3di model) MAP_EXPORT_TEMPLATE_CODE = 4 (workflow for map's export)

The range of template's code 0 - 50 area reserved for auto workflows.

Upload/download wateren- and kerigne-sahpes

Create a symbolic link BUILDOUT_DIR/var/ror_export to the mounted directory (see ROR_KERINGEN_PATH in settings.py):

$ ln -s /mnt/flooding/Flooding/ror_keringen var

Symlink a buildout configuration

Initially, there's no buildout.cfg. You need to make that a symlink to the correct configuration. On your development machine, that is development.cfg (and staging.cfg or production.cfg, for instance on the server):

$> ln -s development.cfg buildout.cfg

Raster Server

We also use an instance of the "raster-server" to serve WMS layers for grid data. The grid data is stored as gislib "pyramids".

To use gislib and raster-server in Flooding, both need to be checked out as development packages, using the "flooding-branch" branch.

Running Buildout, a configuration file for the raster-server is created as etc/rasterserver.json. It says that the rasters are served from BUILDOUT_DIR/var/pyramids. It is possible to symlink /mnt/flooding/Flooding/pyramids to that directory, or to copy a few rasters from the mounted share to that directory, or to change the etc/rasterserver.json.in input file to use something file (in that case, don't commit it).

The command to run the raster-server in development is, in the buildout directory:

$> RASTER_SERVER_SETTINGS=etc/rasterserver.json bin/runflask

The server will run at 0.0.0.0:5000 and visiting it should show a working demo page where the available layers can be shown (although there might be way too many for the page to render if you are using the full Flooding share).

The URL used to find the WMS server is set in the Django settings as RASTER_SERVER_URL. developmentsettings.py sets it to 'http://127.0.0.1:5000/wms' by default, change it to whatever you need in localsettings.py if you are using virtual machines or similar.

Windows (task-server)

• Check out the windows subdirectory, and customize it if needed.
• Check out the objectenbeheer/settings/windows.py module, and customize it if needed.
• Run build_windows.sh from Linux to wrap everything in a nice zip.
• In Windows, download Python 2.7.x from http://www.python.org/download/.
• In Windows, download Psycopg2 from http://www.stickpeople.com/projects/python/win-psycopg/.
• Extract the zip in the configured place, e.g. D:\Programs\flooding.
• In Windows, configure your PYTHONPATH environment variable to point to the absolute path of the flooding\lib subdirectory. If you don't know how to do this, read https://kb.wisc.edu/cae/page.php?id=24500.
• To tune local settings like the database connection, create or edit objectenbeheer\lib\flooding\localsettings.py.