Even though flattr4j offers a number of classes for accessing the Flattr API, internally it all boils down to a few classes that are used for the actual communication. This article gives a peek behind the curtain. It will show you how to fetch a Thing object by its ID, by using the communication classes directly. However, it is not the recommended way to use flattr4j. Also, it requires some in-depth knowlege of the Flattr API.
The communication with the Flattr server is handled by exchanging JSON objects via HTTP. At first, we need a FlattrConnector instance for creating connections to the server:
FlattrConnector connector = new FlattrConnector();
If an AccessToken is required for the call, it needs to be set in a second step:
FlattrConnector connector = new FlattrConnector(); connector.setAccessToken(accessToken);
For open calls (like our call to get a Thing), the access token is not required, but it would not hurt to set it anyways.
Once we have a FlattrConnector, it can be reused for as many calls as we like to. For each call, a new Connection instance is created by the FlattrConnector. At this stage, we have to decide which HTTP method is required. This is documented in the Flattr API documentation. The HTTP method is passed as RequestType parameter to create(RequestType type). For the GET method (which is used most of the time), a convenience method (create()) is also available. The generated Connection can be used for one API call only.
For fetching a Thing, we could use the create() method, but for the sake of this example, let’s use the RequestType instead:
Connection conn = connector.create(RequestType.GET);
The documentation says that for getting a Thing, the URL https://api.flattr.com/rest/v2/things/:id needs to be called, where :id is a placeholder for the thing ID. The base URL is already set in the FlattrConnector, so we only need to set a relative URL in the Connection:
conn.call("things/:id");
Note that the placeholder (:id) is explicitly set in the call string. The corresponding value is set in the next step. flattr4j does the job of resolving the placeholders for us.
conn.parameter("id", "164751");
We’re done! We set up the URL and the parameter. Now it’s time to place the actual call against the Flattr server. As we are only expecting a single result (a single Thing resource), singleResult() is invoked. It sends the request to the server, parses the JSON response and creates a FlattrObject for the result. If an error occures, a FlattrException is thrown instead.
FlattrObject fo = conn.singleResult();
If we would expect multiple results (e.g. a list of Things), we’d invoke result() instead, which returns a Collection of FlattrObject. It is safe to use result() all the time. For API calls that return a single resource, it would still return a collection of a single entry. However it is more convenient to use singleResult() if only one result object is expected. singleResult() would also safely fail if more than one entry or nothing was returned.
The FlattrObject is basically a wrapper around the JSON result. It gives access to the properties within the JSON structure. Let’s say we want to read the URL of the returned Thing. The Flattr REST API documentation says it is stored in a property called (not much surprising) “url”:
String url = fo.get("url"); System.out.println("Thing's URL is: " + url);
And that’s it! We just placed a call against the Flattr REST API and read the result.
The code example above works, but it is a little difficult to read. Most of the Connection methods return itself, so it is possible to chain multiple calls, like in a builder. The example above could also be written like this:
FlattrConnector connector = new FlattrConnector(); FlattrObject fo = connector.create(RequestType.GET) .call("things/:id") .parameter("id", "164751") .singleResult();
FlattrObject gives access to the properties in the JSON result. But it’s much easier to use flattr4j’s Thing instead. Luckily, it takes the FlattrObject as constructor parameter:
Thing thing = new Thing(fo); String url = thing.getUrl(); System.out.println("Thing's URL is: " + url);
The Connection class is able to place all types of calls to the Flattr server (even the OAuth ones, which are a little magic).
For example, this code sends a Submission to the Flattr server. The submission is converted to JSON and sent in the body of a POST request. The returned JSON object contains the ID of the Thing that was created, which is read and displayed.
Submission submission = // your Submission to be sent FlattrConnector connector = new FlattrConnector(); FlattrObject fo = connector.create(RequestType.POST) .call("things") .data(submission.toFlattrObject()) .singleResult(); System.out.println("The submission got Thing ID " + fo.get("id"));
To actually place a call, either singleResult() or result() must be invoked. However, some calls return no result. To check if they succeeded, just invoke result() and ignore the returned value. The next code example deletes a Thing. result() is only invoked to place the actual call and check if the server responded with an error. singleResult() cannot be used here, because it expects exactly one result resource.
FlattrConnector connector = new FlattrConnector(); connector.create(RequestType.DELETE) .call("things/:id") .parameter("id", "1234567") .result();
In order to avoid too much load, Flattr limits the number of calls that can be placed in a certain timespan. The maximum number of calls, and the number of calls that are left, are returned in special HTTP headers. To read them, pass a RateLimit instance to the Connection:
RateLimit lastRateLimit = new RateLimit(); Connection conn = connector.create() .rateLimit(lastRateLimit); // now set the other Connection parameters, place the call and read the result System.out.println("Maximum number of calls: " + lastRateLimit.getLimit()); System.out.println("Number of calls left: " + lastRateLimit.getRemaining());
A RateLimit instance can be reused. It is set by every invocation of the result() or singleResult() methods.