Recently I was using RestSharp to consume my Restful Resouce. and expected exchanging data with JSon between server and client. Below is my C# code.
var client = new RestSharp.RestClient();
var request = new RestRequest(sUrl,Method.POST);
request.RequestFormat = DataFormat.Json;
request.Timeout = TIME_OUT_MILLISECONTS ;
request.AddHeader("Content-Type", "application/json");
request.AddBody(new { appID = sAppId, loginName = sUserName, password=sPassword });
var response = client.Execute(request);
string s=response.Content;//It is always XML format.
The result is not what I expected for(Json data format), although I had set the RequestFormat Json and add Http header Content-Type. So I decided to use the .Net Reflector to found out What happened in the RestClient.Execute method. Here is the code of the method.
public RestClient()
{
...
this.AddHandler("application/json", new JsonDeserializer());
this.AddHandler("application/xml", new XmlDeserializer());
this.AddHandler("text/json", new JsonDeserializer());
this.AddHandler("text/x-json", new JsonDeserializer());
this.AddHandler("text/javascript", new JsonDeserializer());
this.AddHandler("text/xml", new XmlDeserializer());
this.AddHandler("*", new XmlDeserializer());
...
}
I have some questions about it:
As the RestClient adds many kinds of Content-Type into the HttpWebRequest. Is it right way to build a Request? And I think Maybe that is the reason why Response.Content always XML.
I don't know why the RestClient needs to build a HttpWebRequest like that. Any meaning to do that?
If we specified both JSon and XMl message format in a Http Request, which one works finally? Is it allowed?
Thanks. Have a good day.
RestSharp will use the correct handler based on the content type of the response. That's what those AddHandlers are doing; its configuring the RestClient to accept certain content types in the response and mapping those types to deserializers. Normally you would want to set an accept header for the json content type which notifies the server to send json in the response.
request.AddHeader("Accept", "application/json")
Of course, this assumes that the server you are hitting is configured to respond with json.
Related
I have to implement a Vertx POST request. Via Postman, the request is done as shown in the following picture:
The tricky part is that the server expects the key "upgrade_file" for the body. I could not find out how to do that with Vertx. This is what I have so far:
Buffer bodyBuffer = Buffer.buffer(body); // body is byte[]
HttpClientRequest request = ...
request.handler( response -> ...
request.end(bodyBuffer);
How can I set "upgrade_file" as the key for the body?
Use a WebClient instead of the HTTP client, it provides dedicated support for submitting forms.
WebClient client = WebClient.create(vertx);
or if you already have created an http client:
WebClient client = WebClient.wrap(httpClient);
Then create the form data as map and send the form using a the right content type
MultiMap form = MultiMap.caseInsensitiveMultiMap();
form.set("upgrade_file", "...");
// Submit the form as a multipart form body
client.post(8080, "yourhost", "/some_address")
.putHeader("content-type", "multipart/form-data")
.sendForm(form, ar -> {
//do something with the response
});
More example see https://vertx.io/docs/vertx-web-client/java/
If you want to send files, the simplest way is using the sendMultipartForm method of Vertx WebClient.
Create a Multipartform firstly.
MultipartForm form = MultipartForm.create()
.attribute("imageDescription", "a very nice image")
.binaryFileUpload(
"imageFile",
"image.jpg",
"/path/to/image",
"image/jpeg");
Then invoke WebClient.sendMultipartForm(form) to send the request.
The generic Vertx HttpClient is a low-level API, you should searialize the form data into a string or buffer, the format is similar to this example file in Vertx GraphQL testing codes. Then send the buffer to the server side.
Can I return a file with my get request? I want to return a word document to calling angularJS service through REST GET method. Not sure if it is even possible.
You're a bit light on detail, so I'm gonna be a bit light on answer.
A REST request is just... a request. The REST side of things is more the way URLs are defined that what actually happens in the request process itself, which is all still vanilla HTTP.
So, same as with any GET request, if you want to return binary data, set the headers appropriately (<cfheader>), then return the file (<cfcontent>).
So this is how I did it, luckily I got this:
http://smithcustomdev.com/2010/10/20/download-file-to-browser-using-cfscript/
All I have to do was make the method remote and listen to REST service
<cfscript>
private void function serveFile(string filePath){
var fileContent = fileRead(expandPath(filePath));
var context = getPageContext();
context.setFlushOutput(false);
var response = context.getResponse().getResponse();
response.reset();
response.setContentType("text/csv");
response.setContentLength(len(fileContent));
response.setHeader("Content-Disposition","attachment; filename=#listLast(filePath,'\')#");
var out = response.getOutputStream();
out.write(ToBinary(ToBase64(fileContent)));
out.flush();
out.close();
}
</cfscript>
I have a Web API service that I'm trying to access via the console using RestSharp. My RestSharp code looks like this:
RestClient client = new RestClient(baseUrlString);
RestRequest request = new RestRequest("controllername/actionname");
request.RequestFormat = DataFormat.Json;
ProcessQuestion model = new ProcessQuestion()
{
Id = "123456",
InstanceId = "123",
UniqueId = "bfb16a18-d0d6-46ab-a5b3-2f0ebbfe8626",
PostedAnswer = new Dictionary<string, string>() { { "question_7907_1", "selected" }, { "question_7907_2", "selected" } }
};
request.AddBody(model);
var response = client.Execute(request)
My Web API action takes a model that has the same parameters as the above model. When the call executes, the binding fails and the parameter is null. I suspect this is due to the RestRequest.AddBody method prepending application/json to the body value as shown below:
{application/json={"Id":"123456","InstanceId":"123","UniqueId":"bfb16a18-d0d6-46ab-a5b3-2f0ebbfe8626","PostedAnswer":{"question_7907_1":"selected","question_7907_2":"selected"}}}
If I post using Fiddler the body binds to the model properly. Below is the body value I provided in Fiddler:
{'Id':'123456','InstanceId':'123','Uniqueid':'bfb16a18-d0d6-46ab-a5b3-2f0ebbfe8626','PostedAnswer':{'question_7907_1':'selected','question_7907_2':'selected'}}
Note that the body value in fiddler is the same with the exception of prepending the application/json key.
Also to note: I've tried what seems like everything...I've separated the parameters out in the action, used FromBody and FromUri attributes, tried custom DictionaryModelBinder's, tried custom ValueBinders, tried changing the way I'm using RestSharp (AddParameter with a RequestBody parameter, AddObject, different URL styles, etc.).
Has anyone else encountered this, and if so, did you solve it? I chose Web API for this service with hopes the model binding would work as it does in MVC, but I'm seeing that isn't the case.
Thanks
EDIT (resolved):
RestSharp automatically uses the JsonSerializer for objects passed in the AddBody method. I figured I was missing something simple, and indeed I was... Adding the Method.Post parameter to the RestRequest instantiation solved the problem.
Specify the method when creating the request:
RestRequest request = new RestRequest("controllername/actionname", Method.POST);
Not sure what the default serializer is for body - you can try making it explicit:
request.AddBody(request.JsonSerializer.Serialize(model));
I'm not sure where the 'application/json' is coming from - that's the Content-Type header you should be sending with your request, definitely not part of the body. So do this instead:
request.AddHeader("Content-type", "application/json; charset=utf-8");
If this doesn't help, try making your code as similar to the example on their site as possible. Try removing complexity (even if it means changing the required functionality) - get it to a point when it works and build additional functionality on that.
http://restsharp.org/
I am trying to consume a binary stream from a ServiceStack service, as described here:
How to consume a file with a ServiceStack client
I realize that I must use a custom WebClient, since I want direct access to the response stream. However, I would still like to avoid making this client by hand. Instead, I'd like to write something like,
var webClient = new JsonServiceClient(baseUrl)
.ConfigureWebClient(new MyRequestDto { Foo = "bar" }));
This way, I wouldn't have to assemble the URL and query string by hand; and when I change my request DTO, I wouldn't have to remember to change my custom WebClient setup code, either.
Is there a way to accomplish this, somehow ? I've looked at ServiceClientBase.PrepareWebRequest(...), and it does a whole lot of useful stuff that I don't feel like copy/pasting into my own code. I'd love to inherit ServiceClientBase and call that method directly, but it's private, so I can't. Anyone got any other ideas ?
All of ServiceStack's C# Service Clients have both a Global and a Local Request Filter allowing you to initialise the Request and Global and Local Response Filters allowing fine-grained access to the returned HttpWebResponse.
Initializing the WebRequest of all ServiceClients using a Global Request Filter:
ServiceClientBase.HttpWebRequestFilter = httpReq => ConfigureWebClient(httpReq);
Using the Local Request Filter:
var client = new JsonServiceClient(baseUrl) {
LocalHttpWebRequestFilter = httpReq => ConfigureWebClient(httpReq)
};
Although if you just want the binary response All Service Clients allow you to specify either a string, byte[], Stream or HttpWebResponse as your Generic Response type and it will return what was requested. See the Service Client wiki page for more examples of these.
Here's how you can retrieve binary responses:
byte[] responseBytes = client.Get<byte[]>("/poco/World");
var dto = responseBytes.FromUtf8Bytes().FromJson<PocoResponse>();
Or with a Stream:
using (Stream responseStream = client.Get<Stream>("/poco/World")) {
var dto = responseStream.ReadFully().FromUtf8Bytes().FromJson<PocoResponse>();
}
Or from the underlying HttpWebResponse:
HttpWebResponse webResponse = client.Get<HttpWebResponse>("/poco/World");
using (var stream = webResponse.GetResponseStream())
using (var sr = new StreamReader(stream)) {
var dto = sr.ReadToEnd().FromJson<PocoResponse>();
}
You can change Open Source Software, you don't have to work around it
ServiceStack is Open Source software, there's no need to look for hacks or try to work around something that's not accessible or not available, make a pull-request to change what you want and if it's a valid request for a common use-case it will likely be accepted. Otherwise feel free to take a fork of the Source code and customize it as you wish.
I am developing a new REST-full webservice for our application,I wanted to send the reqest data in requestHeader instead of sending as query param, as my request data is large.
I have My jquery code like below to add json to the request header and call the REST service GET method.
$.ajax({
beforeSend: function(req) {
req.setRequestHeader("test", "{name:mouli, id:918}");},
type : "GET",
data :'',
dataType : "jsonp",
url : 'http://localhost:29801/RestFulJSONExample/rest/jobdesc?callback=?',
success : function(data) {
alert("invoked");
}
});
});
And my GET method in my REST service is like
#GET
#Produces("application/javascript")
public JSONWithPadding getJobDescription(#Context HttpHeaders headers) {
List<String> requestHeader = headers.getRequestHeader("test");
//some logic here.
}
i could able to get the JSON object from the request header which i have added in the jquery request.
My question is ..
Can i follow this approach? is it secure and safe?
If not please tell me the other way?
What appears at the right side of the ":" in a header is mostly free. You have to take into account character set restriction in HTTP, and possible carriage returns in the JSON value (you know, headers of more than one line have a specific syntax). If your JSON examples are relatively simple, then I see no problem in that. It is just another way of organizing the actual value of a header line.