Global Interceptor

Global request interceptor

Usually, we need to use the same configuration for all requests, such as adding token and timestamp to the request header. At this time, we can set a global request interceptor, which will be triggered before all requests. We can set this interceptor Set request parameters uniformly.

const alovaInstance = createAlova({
  // ...
  //The function parameter is a method instance, including request data such as url, params, data, headers, etc.
  // You can modify these data freely
  beforeRequest(method) {
    // Suppose we need to add token to the request header
    method.config.headers.token = 'token';
  }
});

You can also set beforeRequest as an asynchronous function.

const alovaInstance = createAlova({
  // ...
  async beforeRequest(method) {
    //Perform some asynchronous tasks
    // ...
  }
});

Global response interceptor

When we want to uniformly parse response data, uniformly handle errors, and uniformly handle request completion, we can specify a global response interceptor when creating an alova instance. The response interceptor includes an interceptor for successful requests and an interceptor for failed requests. and request completion interceptors.

const alovaInstance = createAlova({
  // ...
  // Use responded object to specify the interceptor for successful request and the interceptor for failed request respectively.
  responded: {
    // Interceptor for successful request
    // When using the GlobalFetch request adapter, the first parameter receives the Response object
    // The second parameter is the method instance of the current request. You can use it to synchronize the configuration information before and after the request.
    onSuccess: async (response, method) => {
      if (response.status >= 400) {
        throw new Error(response.statusText);
      }
      const json = await response.json();
      if (json.code !== 200) {
        // This request will throw an error when an error is thrown or a Promise instance with reject status is returned.
        throw new Error(json.message);
      }

      //The parsed response data will be passed to the transformData hook function of the method instance. These functions will be explained later.
      return json.data;
    },

    // Interceptor for request failure
    // This interceptor will be entered when a request error occurs.
    // The second parameter is the method instance of the current request. You can use it to synchronize the configuration information before and after the request.
    onError: (err, method) => {
      alert(error.message);
    },

    //Interceptor for request completion
    // When you need to execute logic whether the request succeeds, fails, or hits the cache, you can specify a global `onComplete` interceptor when creating an alova instance, such as turning off the request loading state.
    // Receive the method instance of the current request
    onComplete: async method => {
      // Process request completion logic
    }
  }
});

If you do not need to set an interceptor for request failure or completion, you can directly pass in the interceptor function for successful request, instead of setting a callback through an object.

const alovaInstance = createAlova({
  // ...
  async responded(response, method) {
    // Interceptor for successful request
  }
});

Interceptor triggering instructions

When you use GlobalFetch to request the adapter, due to the characteristics of window.fetch, the onError interceptor will only be triggered when the connection times out or the connection is aborted. In other cases, the onSuccess interceptor will be triggered. For details, please Check here

Special attention

1. onSuccess, onError and onComplete can be set as synchronous functions and asynchronous functions.
2. The onError callback is a capture function for request errors. An error thrown in onSuccess will not trigger onError. When an error is caught but no error is thrown or a Promise instance that returns reject status is used, the request will be considered successful and no response data will be obtained.
3. In 2.0.x and previous versions, responded was incorrectly spelled as responsed. In 2.1.0, the two have been made compatible. It is recommended to use responded instead of responsed in subsequent versions. .