Latest Contributions by KevinHibma

Simply by looking at the picture of the model, it looks correct. I would have personally used copy features at the end instead of MakeFeatureLayer....but I doubt that'd cause you any issues. The quick test here is - consume the service back inside ArcMap. If this works correctly every time, you know the service is good and the issue is related to the WAB configuration/use of the GPService. My guess at the issue: Its the WAB. Or the WAB you're using? Are you using a local developer version? If I take your GP Service URL and plug it into the ArcGIS Online WAB - the output from your service shows as I'd expect it to when configuring the widget. That is, the output is a points layer and I just need to set the drawing options. From your screen shot you see items like min/max scale as well as target layer.My WAB doesn't show me this. I'm not sure how or why you've got this, but to me that it seems like your configuration thinks the output of the GP Service is something else (not just a simple points layer output). ... View more

How are you referencing the executable? If you've got a full path variable:  myExe = r'c:\folder\doSomething.exe'  try swapping it to something like: myFolder = r'c:\folder'  #this is the same item thats registered with the datastore myExe = os.path.join(myFolder, 'doSomething.exe') Or maybe you have it that way, and you want to switch it to the full path. I'm not in a position to test this right now, but I know one (actually though, both ways) should work. What version of Server? ... View more

No it wont. The task will only make the output zip file. I just assumed because you were using doing a variable number of input layers to clip that your end goal was to actually extract them for download. You're going to need to look at using the iterators inside model builder and build the task yourself. However, at this point I'd say a python script might be more straight forward and give you more flexibility. I'm not entirely sure you can setup a model with an iterator to accept multiple layers and iterator over them. (Because your input will probably be featurelayer:multivalue). Thus in a script you'd accept the multivalue, then split the inputs and iterate over them in a loop. I'll point this out though - GP Services do not like when you have variable numbers of outputs. For example, if your tool accepts 1-5 layers as an input, you could set it up to have 1-5 output featureclasses (after the buffer/clip operation). The problem being, GP Services want well defined scenarios, so if your user only selects 2 layers, but the service expects 5, it doesnt get all the outputs back it was expecting too. So you'll need to come up with a single parameter, passing back a multivalue. (I cant remember off the top of my head if this will work, its been a long time since I've worked with output:multivalue).  The take home point here is:  you'll have to set your model or script up in a given way to work properly as a gp service and I can't tell you that particular way without knowing your exact requirements. However, this all could be moot if you're just wanting to return the "clip" results as a map image (instead of the features sent back to the client to be drawn). That is a snap shot of the output to be drawn on top of your map. But at this point I'm making a lot of assumptions on your workflow and end goals. Hopefully some of the thoughts above will allow you to move forward with your workflow. ... View more

Look at using the Extract Data Task​. It already does what you want (clips multiple layers). The only thing you want to do is pull that model into your own model where you'll have buffer as the first tool and the output of buffer is passed as the selecting features to the extract data task. (Because the EDT is setup to accept a polygon that a user could draw, thus you're just tweaking the front end experience) ... View more

Has anyone tried Project Templates? Start a new project. Add a map or scene, regularly used folder connections or databases and anything else you want direct access to. Next time you make a new project, don't use a blank one, select the template you previously made and give your new project a name/location. You should start with all the items you setup from before. More information here. Please don't consider this a dismissal of the original idea/ enhancement, I'm just offering alternatives that can move you towards your goals, today. ... View more

You say widget, so I assume you're using the WebAppBuilder (WAB)? If so, yes, it is smart enough. If your input parameter is setup as an interactive feature set (like you draw a box inside ArcMap), then when you use the WAB - GP widget against the GP Service you've made, it is smart enough to provide an interactive draw option in the web app itself. So no extra code required. ... View more

You're correct that GP Server considers CAD as a non-transportable type. You'll need to modify your workflow into a script. From there you can either use a FILE or a STRING variable as input. (Sounds more like you want to use STRING). Within the script you're responsible for passing the "cad file path" to the CAD to Geodatabase tool. (I just tried in MB but the CAD to Geo tool seems pretty strict on the inputs, thus you wont be able to make a multi-string variable or a multi-file variable and hook it up to the point the tool will run. A script ​should​ get you around  the validation requirements of the tool.) ... View more

Ok, your 2nd problem. GP Services are very strict with their input and output params. Optional output params in a GP Service is well........ not straight forward. Outputs get hardcoded with a gp service. When you run your tool in ArcMap before publishing, this sets up the "footprint" if you will of what the service will be like. So if you had 3 outputs when you ran in AM, once you publish it the service will want 3 outputs. Not 2, not 4, but 3. So anytime you start playing the optional game you're going to get into trouble. For the 1st issue - I dont really understand what you're saying here. I can only guess this is a "tool problem", that is how the tool has been setup/configured. You're having the problem here before publishing? Or this is the published gp service? You're expecting to get the result back from the gp service and it gets added to ArcMap? As long as the output is a featureclass in the tool, yes it should be automatically added. ... View more

What is your end goal? Using arcpy.mapping.MapDocument has pretty limited use in a GP Service. Its mainly good for printing and exporting. If you're trying to use this workflow to update a map you have locally...well I dont see the connection. If you can explain what you want to do (not how you're trying to do it), maybe someone can suggest a workflow. ... View more

You must be using the 10.0 version of the tutorial. Try the 10.1+ version here:   Geoprocessing service example: Clip And Ship—Sharing geoprocessing workflows | ArcGIS for Desktop (The steps you mention are telling you to make a "tool layer", a concept that does not exist in 10.1+) ... View more

I'm sorry to report that we did not fix this in 10.3.1. We had attempted a fix but were not happy with the results, as such the issue still present. We are committed to fixing this issue and will try to it addressed in the next release (10.4) ... View more

Without seeing your model or script, it sounds like the input parameter to it is a file (GPDataFile as you reference). Before publishing you'd need to setup the model or script with an input parameter of TABLE. Take special note though, that when publishing the result as a service, depending how you provided input to the parameter it may make a best guess and try to change it. Ensure that the input to your service stays as a "user defined value" (it'll show the schema of the table in the service editor for you to confirm). Once published, it'll be a GPRecordSet. Getting the required input JSON correct can be troublesome. If you go into ArcMap and open the Python window, you can get JSON code from your CSV by putting these lines: rs = arcpy.RecordSet() rs.load("c:/path/to/file.csv") rs.JSON ... View more