This is a way to bring at least part of QGIS Processing Functionality to World Wide Web Through WPS.

You will install pywps the quick and dirty way :

1. cd to PyWPS-QGIS-Processing directory:

   $ cd /home/foo/bar/PyWPS-QGIS-Processing

2. Simply copy PyWPS directory to the target directory:

   $ cp PyWPS /usr/local/

We used an embedded PyWPS because we updated it to initialize a QGIS environment for QGIS-Processing and to use QGIS-Server for as reference outputs

For more information about PyWPS : http://pywps.wald.intevation.org/documentation/

PyWPS-QGIS-Processing comes with :

1. pywps.cgi
2. pywps.cfg
3. processes

Update pywps.cgi to adapt it to your PyWPS and QGIS install :

• DISPLAY, you need to create a virtual display for example with Xvfd
• PYTHONPATH to your QGIS install and the plugins directories
• LD_LIBRARY_PATH to your QGIS install
• PYWPS_CFG to the path where you put the PyWPS-QGIS-Processing pywps.cfg
• PYWPS_PROCESSES to the directory path where you put PyWPS-QGIS-Processing processes directory
• the path to the PyWPS wps python script

Update pywps.cfg like it's describe in PyWPS install.

We add two section to pywps.cfg to configure PyWPS-QGIS-Processing :

• [qgis]

  • prefix path to your QGIS installation. The QGIS installation directory contains lib and share directory
  • user_folder path to the folder where QGIS Processing will generate log and temporary files. For QGIS Desktop, this directory is /home/user/.qgis2
  • providers the providers list separated by comma you want to publish. The list can contain qgis,gdalogr,script,model,r,grass,grass70,saga. If the list is empty all providers will be published
  • algs_filter text to filter algorithms. Text will be searched in algorithm name and commandLineName
  • algs an algorithms white list separated by comma you want to publish. Algorithms are defined by their commandLineName, for example qgis:voronoipolygons,qgis:fixeddistancebuffer,gdalogr:aspect,gdalogr:hillshade,gdalogr:roughness,gdalogr:slope,gdalogr:contour
  • projects_folder path to the directory which contains qgis projects. You can defined a project by algorithms to provide complex input data.
  • qgisserveraddress the QGIS-Server address to publish outputs as webservices

• [qgis_processing]

  • the ACTIVE*_ activate providers, you can activate providers without publishing them if they are needed in models
  • the _*FOLDER for the pathes to your models, scripts and R scripts
  • the SAGA*_ are for the SAGA provider configuration
  • all the providers configuration parameters can be done here.

• Copy :

  • models form your QGIS Processing models to the [qgis_processing] MODELS_FOLDER
  • scripts form your QGIS Processing scripts to the [qgis_processing] SCRIPTS_FOLDER
  • rscripts from your R QGIS Processing to the [qgis_processing] R_FOLDER

By default PyWPS-QGIS-Processing provide a model modeler:buffer which is a simplification of the process qgis:fixeddistancebuffer

Once all is conform to your PyWPS and QGIS installation you can open the links :

• http://localhost/cgi-bin/pywps.cgi?SERVICE=WPS&REQUEST=GetCapabilities
• http://localhost/cgi-bin/pywps.cgi?SERVICE=WPS&VERSION=1.0.0&REQUEST=describeprocess&IDENTIFIER=modeler:buffer
• http://localhost/cgi-bin/pywps.cgi?SERVICE=WPS&VERSION=1.0.0&REQUEST=execute&IDENTIFIER=modeler:buffer&DATAINPUTS=\[NUMBER_DIST=20.0;VECTORLAYER_BASE=http://apps.esdi-humboldt.cz/classification/traning_areas/training_areas_en.gml]