I have a C# class that does an address lookup. I want to expose this as an Azure function. I've been going through the documentation but can't see how I can/if it's possible to do the following:
I have a Git repository in Team Services that contains a class library of my AddressLookup. Can my Function reference this project?
If I look at the folder structure of the site I can see it has copied over all the source files from the Git repository, can I get it to build the solution or does it literally just pull all the files?
Where in the solution do I put the function? Do I create a solution folder of the name of the function and place the relevant files in there?
My AddressLookup class returns an object that is defined in the class library. Will the function be able to use and return this?
Thanks
Alex
Follow-up to Q1: Are you trying to setup CI? For continuous integration with Azure Functions, you may reference the following:
https://azure.microsoft.com/en-us/documentation/articles/functions-continuous-deployment/#setting-up-continuous-deployment
http://flgmwt.github.io/azure/azure-functions/c-sharp/2016/04/04/azure-fns-with-ci.html
--Update 10/17--
Specific to Team Services, here are the steps:
Make sure that your VSTS account is linked to your Azure Subscription. Follow the instructions in this article.
Navigate to the Functions Portal for your Function App and click on Function app Settings -> Configure continuous integration.
In the Deployements blade, click on Setup and configure your Deployment source information (see sample snapshot below). Click on the OK button. Wait for the sync to succeed. Close the Deployments blade.
Give it a minute and refresh your Functions Portal session. You should now see the function added to your Function site. The snapshot below is my AddressLookup function that was synced from my Team Services project named MyFirstProject.
Note the disclaimer message above the Code editor. If you hook up CI for your Function, you will not be able to edit it in the Functions Portal. Since this particular example requires a request body, you will need to test it using Postman.
--End of update 10/17--
Answer to Q2:
Here's a good documentation describing the folder structure of Azure Functions:
https://azure.microsoft.com/en-us/documentation/articles/functions-reference/
I also recommend the follow-up documentation specific to C# development for Azure Functions:
https://azure.microsoft.com/en-us/documentation/articles/functions-reference-csharp/
Answer to Q3 & Q4: I will attempt to answer these by providing a sample implementation. I don't have any context on the implementation of your AddressLookup library, however, in the interest of providing an example, I am going to take a wild leap and assume that it is a library that will perform some Geocoding operations. Assuming again that you want to use this library in an HTTP-triggered Function, you may begin by first generating the AddressLookup.dll and then uploading it to the bin folder inside your Function. You may then reference that DLL from your Function script.
For instance, using this article as a reference, I generated a AddressLookup.dll library in Visual Studio that has the following implementation. This DLL will serve as a proxy for your AddressLookup library so that I can demonstrate how we can use it in a Function.
using System;
using System.Globalization;
using System.IO;
using System.Linq;
using System.Net;
using System.Web;
using System.Xml.Linq;
namespace AddressLookup
{
public class GeoLocation
{
public double Longitude { get; set; }
public double Latitude { get; set; }
}
public class GeoCoder
{
private const string geoCodeLookupUrlPattern =
"https://maps.googleapis.com/maps/api/geocode/xml?address={0}&key={1}";
private const string addressLookupUrlPattern =
"https://maps.googleapis.com/maps/api/geocode/xml?latlng={0},{1}&key={2}";
private string _apiKey = null;
public GeoCoder(string apiKey)
{
if (string.IsNullOrEmpty(apiKey))
{
throw new ArgumentNullException("apiKey");
}
_apiKey = apiKey;
}
public GeoLocation GetGeoLocation(string address)
{
GeoLocation loc = null;
string encodedAddress = HttpUtility.UrlEncode(address);
string url = string.Format(geoCodeLookupUrlPattern, encodedAddress, _apiKey);
WebRequest request = WebRequest.Create(url);
using (WebResponse response = request.GetResponse())
{
using (Stream stream = response.GetResponseStream())
{
if (stream != null)
{
XDocument document = XDocument.Load(new StreamReader(stream));
XElement longitudeElement = document.Descendants("lng").FirstOrDefault();
XElement latitudeElement = document.Descendants("lat").FirstOrDefault();
if (longitudeElement != null && latitudeElement != null)
{
loc = new GeoLocation
{
Longitude = Double.Parse(longitudeElement.Value, CultureInfo.InvariantCulture),
Latitude = Double.Parse(latitudeElement.Value, CultureInfo.InvariantCulture)
};
}
}
}
}
return loc;
}
public string GetAddress(GeoLocation loc)
{
string address = null;
string url = string.Format(addressLookupUrlPattern, loc.Latitude, loc.Longitude, _apiKey);
WebRequest request = WebRequest.Create(url);
using (WebResponse response = request.GetResponse())
{
using (Stream stream = response.GetResponseStream())
{
if (stream != null)
{
XDocument document = XDocument.Load(new StreamReader(stream));
XElement element = document.Descendants("formatted_address").FirstOrDefault();
if (element != null)
{
address = element.Value;
}
}
}
}
return address;
}
}
}
Now, let's create an HTTP-Triggered Function by performing the following steps:
Go to the Functions Portal. Create a Function using the HTTP
Trigger - C# template.
Fill in the name (e.g., AddressLookup) and authorization level (e.g., Anonymous). You should now see a Function named AddressLookup created with some pre-populated code.
On the left pane, click on the Function app settings button.
Optional: Click on Configure app Settings. Under the "App settings" section, add a value for the key GoogleMapsAPIKey with your api key, then click on the Save button. Note: If you skip this step, then you will need to hard-code the key in your function code later.
Next, use the Kudu console to upload your DLL. Click on the Go to Kudu button. This will launch a new browser window
with a cmd console. Type the following to navigate to your
Function directory,
cd site\wwwroot\AddressLookup
Create a bin folder by typing mkdir bin at the command prompt as follows,
Double-click on the bin folder and upload (see "Add files") the AddressLookup.dll into the folder. When you are done, you should a similar snapshot below,
Go back to the Functions Portal. In your Function's editor, at the bottom of the Code section, click on View Files. You should now see the newly created bin folder as follows,
Replace the contents of the pre-populated Function script with the following code
#r "AddressLookup.dll"
using System;
using AddressLookup;
using System.Net;
public static async Task<HttpResponseMessage> Run(HttpRequestMessage req, TraceWriter log)
{
log.Info($"C# HTTP trigger function processed a request. RequestUri={req.RequestUri}");
// Reading environment variable from App Settings, replace with hardcoded value if not using App settings
string apiKey = System.Environment.GetEnvironmentVariable("GoogleMapsAPIKey", EnvironmentVariableTarget.Process);
// Get request body
dynamic data = await req.Content.ReadAsAsync<object>();
string address = data?.address;
string name = data?.name;
GeoCoder geoCoder = new GeoCoder(apiKey);
GeoLocation loc = geoCoder.GetGeoLocation(address);
string formattedAddress = geoCoder.GetAddress(loc);
HttpResponseMessage message = null;
if (name == null)
{
message = req.CreateResponse(HttpStatusCode.BadRequest, "Please pass a name in the request body");
}
else
{
var msg = $"Hello {name}. Lon: '{loc.Longitude}', Lat: '{loc.Latitude}', Formatted address: '{formattedAddress}'";
message = req.CreateResponse(HttpStatusCode.OK, msg);
}
return message;
}
Click on the Save button.
In the "Run" section, supply the following request body,
{
"name": "Azure",
"address": "One Microsoft Way Redmond WA 98052"
}
Click on the Run button.
You should see some log entries similar to the following,
2016-10-15T03:54:31.538 C# HTTP trigger function processed a request. RequestUri=https://myfunction.azurewebsites.net/api/addresslookup
2016-10-15T03:54:31.773 Function completed (Success, Id=e4308c0f-a615-4d43-8b16-3a6afc017f73)
and the following HTTP response message,
"Hello Azure. Lon: '-122.1283833', Lat: '47.6393225', Formatted address: '1 Microsoft Way, Redmond, WA 98052, USA'"
Since this is a HTTP-triggered Function, you may also test your Function using Postman. See snapshot below,
If you upload your own DLL in step 5 and edit the Function code to call your library, the Function should work just as well.
Is the assembly you want to reference frozen, or do you want to see updates? If you don't want to see updates, see the answer by Ling Toh.
But if you want to see updates when the assembly updates:
Link your function to a some form of continuous delivery. The official documentation explains how to do this. At the moment, you seem to need a separate git repository, VSTS project (or whatever) to do this easily. (It is possible to edit the deployment process in Kudu, but I would avoid that if you possibly can).
The functions project should contain only the functions code themselves. So it should contain something like the followling:
global.json
host.json
packages.config
Web.config
function1/
function1/run.csx
function1/project.json
function1/function.json
Where you should replace function1 by the name of your function.
Once you've configured this to push to your functions host, you're most of the way to where you want to be. Next, add a function1/run subdirectory where you place yourAssembly.dll. This should be automatically copied here on a successful build of the assembly project. I don't have experience of VSTS to know exactly the best way to do this, so you might need to ask another question.
You've now placed the assembly in the right place. You now refer to it by adding the assembly reference line to the top of your run.csx:
#r "yourAssembly.dll"
Note that all of the references have to be before everything else at the top of the run.csx. So you can put it after other references, but it has to be before anythign else, including load specifications.
Note that for custom assemblies you include the .dll and quote the file, whereas for framework assemblies that aren't referenced by default, you don't
In an ideal world, this will be enough.
But functions doesn't trigger a rebuild if the assembly is updated at the moment, only if project.json, run.csx or function.json is updated. So as part of this process, I then look at the end of the file and add or remove a blank line. It needs to be meaningless (so that you're not changing something important), but enough for your version control tool to think that the file has changed. Clearly, this step is not necessary if you've also made other changes to the file.
If you're using git, you'd now commit and push the changes up. Functions will see that function has changed, and recompile. You should see this in the log pane in the function itself.
Note that if you have two functions using the same dependent assembly, you need to copy it into both function folders; this can't be shared.
I am trying to use the SFAutoComplete control from SyncFusion in a Xamarin iPad app. (only iPad).
I am not able to get any sort of change event to fire.
What I've tried:
If you download SyncFusion and install it, it comes with a "SampleBrowser" app that has samples for all the controls in the suite.
If you open that SampleBrowser in visual studio and open the AutoComplete_Tablet.cs file after line 97, I've added this code:
countryAutoComplete.ValueChanged += (sender, args) =>
{
suggestionModeLabel.Text = "IT WORKED!";
};
But it never fires.
I've tried to use several different events from the list of events this control has (partial list from screenshot):
None of them seem to fire (I haven't tried ALL of them).
What do I need to do to get one of these events to fire? What am I missing?
Thanks for using Syncfusion Controls.
Delegate property can be used to hook the SFAutoComplete's events as per in the following code example,
Declaration code for Delegate property
SFAutoComplete autocomplete = new SFAutoComplete();
autocomplete.Delegate = new SFAutoCompleteDelegate();
The way to hook the events in SFAutoComplete
public class SFAutoCompleteDelegate : AutoCompleteDelegate
{
public override void DidTextChange(SFAutoComplete SFAutoComplete, string value)
{
//It fired while changing the text in AutoComplete
}
public override void DidSelectionChange(SFAutoComplete SFAutoComplete, string value)
{
//It fired while changing the suggestion from suggestion box.
}
}
We have created a sample for achieving your requirement. Please download the same from the following link
Link:http://www.syncfusion.com/downloads/support/forum/125261/ze/testingAutoComplete_21799375630
Thanks & Regards,
Hemalatha M.R
How do I do this? If you could please kindly include the code for message map and the function itself, that would be greatly appreciated.
EDIT:
More specifically I am wondering how OnFileSave() links to OnSaveDocument(LPCSTR lpszPathName)
How does OnFileSave get lpszPathName?
You don't need to do anything special to override OnSaveDocument(...) it's already a virtual function in CDocument, so your derived class can just declare virtual BOOL OnSaveDocument(LPCTSTR lpszPathName); in it's header, then implement it in the document. Nothing is needed in the message map. OnSaveDocument will be called by the framework as part of OnFileSave which is a handler in the base class for ID_FILE_SAVE. The lpszPathName refers to m_strPathName when called by OnFileSafe, which is set when opening a file or by calling SetPathName. If it's empty when saving, the user is prompted for a file name.
CDocument::OnFileSave is the message handler for the Save menu command. To handle it yourself put this in your document class message map:
ON_COMMAND(ID_FILE_SAVE, OnFileSave)
and add your function:
void CYOURDOCUMENT::OnFileSave()
{
CDocument::OnFileSave();
}
To see everything it does put a breakpoint in your function and start single-stepping.
I have a basic question. I have developed a Python component that just prints the value of a simple property to the console screen and a waveform that just runs the component.
I'm trying to create a component control panel that displays the current property value (a string) and also allows me to change the value of the property. When I change the value of the property, I would like the value printed to the console window to change too. My understanding is that I need to bind the property to a text box on the control panel.
I attempted to follow the binding example in section 24.4.1 of the documentation (http://redhawksdr.github.io/Documentation/mainch24.html), by adding their bind code to the refresh() function that was automatically generated by the SCA Component Control Panel wizard. However, when I view the control panel, I get the error:
An internal error occurred during: "Refresh Component".
null argument: Validation realm cannot be null
I am fairly new to Java, but my guess is that I'm receiving the error because there was some problem when creating the local copy of the SCA component (variable named "component" in the code below).
Here is my control panel code:
public void createPartControl(final Composite main) {
main.setLayout(new GridLayout(2, false));
propValue = new Text(main, SWT.BORDER);
propValue.setLayoutData(new GridData(SWT.FILL, SWT.CENTER, true, false, 1, 1));
}
public void refresh() {
Job refreshJob = new Job("Refresh Component") {
#Override
protected IStatus run(IProgressMonitor monitor) {
// bind text box to component property value
final ScaComponent component = getInput();
EMFDataBindingContext context = new EMFDataBindingContext();
IObservableValue observable = SCAObservables.observeSimpleProperty(component,"someText");
context.bindValue(WidgetProperties.text().observeDelayed(5000, propValue),observable);
return Status.OK_STATUS;
}
};
refreshJob.setSystem(true);
refreshJob.schedule();
}
For reference, I am running REDHAWK version 1.8.2 on RHEL 5.8.
I believe the problem here is you are trying to create a binding from within a Job.
Creation of bindings should be done within the SWT thread. The error you are getting is indicating that was unable to acquire a realm to execute within. If this code is executed within the SWT thread it will correctly acquire the SWT realm.
Text propValue = new Text(main, SWT.BORDER);
EMFDataBindingContext context = new EMFDataBindingContext();
IObservableValue observable = SCAObservables.observeSimpleProperty(component, "simplePropId");
context.bindValue(WidgetProperties.text().observeDelayed(5000, propValue), observable);
Notice in the example how the bindings are immediately created. This ensures the bindings are within the correct realm.
Control panels will have the input model change occasionally. Therefore it may be necessary to create the bindings within the refresh or setInput methods. These methods are also called from the SWT thread.
Note you should free the old context before rebinding to the new model by using context.dispose and creating a new context.
For an example on how data bindings work checkout: http://www.vogella.com/articles/EclipseDataBinding/article.html
i'm working with VS2012, Windows Phone 8 SDK.
I need to access the camera natively.
I have sucesfully created a PhotoCaptureDevice instance by calling PhotoCaptureDevice::OpenAsync, then get the opened device - mCaptureObject - with the help of create_task() and task.then().
however,
mCaptureObject->Close()
caueses compile error.
CameraStream.cpp(xxx): error C2039: 'Close' : is not a member of Windows::Phone::Media::Capture::PhotoCaptureDevice'
here's a link to msdn:
PhotoCaptureDevice.Close()
i checked the photocapturedevice class declaration:
public : virtual unknown-type Close() new sealed = Platform::IDisposable::Dispose
Member of Windows::Phone::Media::Capture::PhotoCaptureDevice
Summary:
Releases resources that are associated with the capture device.
Any idea?
To call Close() on an object, you should invoke that object's destructor. This will happen naturally when the object no longer has references, or you can call delete mCaptureObject to force the destructor to run immediately. (If you do this while the object still has living references, those other references will be invalid and method calls should throw a DisconnectedException.)