cdatweb | Web visualization framework of UV-CDAT

Installing the django components is relatively simple. It is recommended you aren’t in your UV-CDAT environment for this section. It is also recommended that you run inside a [virtual environment](http://virtualenv.readthedocs.org/en/latest/virtualenv.html) as follows:. You can leave the virtual environment at any time by running deactivate.
Refer to one of the items below for starting up the visualization server. Once the server is running, browse to [http://localhost:8000/vtk/viewer.html](http://localhost:8000/vtk/viewer.html). In case of trouble, see the server logs in tmp/logs. By default, CDATWeb is configured to use a development visualization server at LLNL. In order to use a local visualization server, you will need to modify your django configuration at cdatweb/settings/development.py. The value of VISUALIZATION_LAUNCHER should be set to your launcher service endpoint. By default this is.
Running vtkWeb from a docker image (Mac OS or Linux ony) You first must have [Docker](https://docs.docker.com/) installed. See the installation guide for more information. On OSX, the setup here assumes you have `docker-machine` on you `PATH` and that you have a docker machine named `docker` (see the docker setup instructions below). First, get the latest docker image from dockerhub by running ``` docker pull uvcdat/cdatweb-vtkweb ``` Now go to the `vis_server` directory and install the requirements for the vtkweb launcher ``` pip install -r requirements.txt ``` You should now be able to run the launcher ``` python launcher.py config-docker.json ``` This will serve the launcher at port 7000 on your machine. At this point, you can start up the django server from a different terminal window as described above. If everything is working correctly, you should see a list of files when you browse to [http://localhost:8000/vtk/viewer.html](http://localhost:8000/vtk/viewer.html).
Running a local vtkWeb instance As noted above, you may be able use a prebuilt UV-CDAT, but if you run into troubles, try to rebuild the latest version with the following option: `cmake -D -DCDAT_BUILD_WEB=ON /path/to/uvcdat`. Make sure you are not in your django virtual environment and source your UV-CDAT setup script. Source uvcdat environment: `. install/bin/setup_runtime.sh` and make sure netcdf is built with DAP support by running `nc-config --all` and checking if `--has-dap` is on. If that is not the case you won't be able to open files from `http://test.opendap.org` If everything is setup correctly, you should be able to run the following: ```bash cd vis_server ./start.sh ```
Running vtkWeb on a remote machine The procedure for serving vtkWeb from a remote machine (or a cluster) is similar to running locally; however, you must edit the file `vis_server/config.json`. The details of making this work are beyond the scope of this document. See [ParaViewWeb documentation](http://pvw.kitware.com/#!/guide/python_launcher) for more details.
On Mac OS, docker containers must run inside a light weight virtual machine. There is a command line utility for managing instances of these VM’s called [docker-machine](https://docs.docker.com/machine/). Docker-machine replaces and earlier tool called boot2docker and supports provisioning on to VirtualBox, VMWare, as well as many cloud providers. For local development, the following command will provision a docker host in VirtualBox with the machine name of docker. To use that host in you’re current shell, you need to run. after which you can use docker as normal. For the docker.py script in this repository, it is assumed that the docker host is called “docker”.