Hello All, I am looking for help on setting up a geoprocessing widget for my viewer.
Here are the pieces that I have: I have a functioning model, in ArcGIS, that runs the process that I want. I have all layers that will be used for the model stored in the same file geodatabase.
The widget will be user driven.
So what I would like help on is connecting the dots on how to make such a widget work. I believe that the geodatabase and model need to go up on the server for access. Correct?
What I am not sure is how to collect user input and then pass it to the model and return the result back to the viewer.
Any help on how this all works would be greatly appreciated.
You need to set up your model to run from the server, this means that your server needs to be able to access any data that is used within it (i.e. the data must be on a shared area that the server can see and use UNC paths to the data). Output data must be written to a scratch area, see:
Once you have the model set to run from the server you need to publish it and then test the model again in ArcMap. If you experience problems at this stage when testing your service in ArcMap you need to do some detailed analysis, a good place to start is turn on verbose logging in server manager and then check for any errors. A common one is the number of features being returned (I'm using 9.3.1 which has default value of 500, if this number is exceeded you just do not get a return). If you are happy that your service is working correctly move onto integrating it with the flex viewer, a good place to start is using the out of the box geoprocessing widget:
This handles most geoprocessing tasks by just setting up the config file. If you experience problems at this stage a good way of checking that the correct values are being passed to the server is to use Firefox with firebug so you can see the request being made. Again use server manager log files to make sure there is no errors being generated.
You need to set up your model to run from the server, this means that your server needs to be able to access any data that is used within it (i.e. the data must be on a shared area that the server can see and use UNC paths to the data). Output data must be written to a scratch area, see:
Once you have the model set to run from the server you need to publish it and then test the model again in ArcMap. If you experience problems at this stage when testing your service in ArcMap you need to do some detailed analysis, a good place to start is turn on verbose logging in server manager and then check for any errors. A common one is the number of features being returned (I'm using 9.3.1 which has default value of 500, if this number is exceeded you just do not get a return). If you are happy that your service is working correctly move onto integrating it with the flex viewer, a good place to start is using the out of the box geoprocessing widget:
This handles most geoprocessing tasks by just setting up the config file. If you experience problems at this stage a good way of checking that the correct values are being passed to the server is to use Firefox with firebug so you can see the request being made. Again use server manager log files to make sure there is no errors being generated.
