"fetchff" stands for "fetch fast & flexibly"

Why?

Managing multitude of API connections in large Frontend Application can be complex, time-consuming and hard to scale. fetchff simplifies the process by offering a simple, declarative approach to API handling using Repository Pattern. It reduces the need for extensive setup, middlewares, retries, custom caching, and heavy plugins, and lets developers focus on data handling and application logic.

Key Benefits:

✅ Small: Minimal code footprint of ~3KB gzipped for managing extensive APIs.

✅ Immutable: Every request has its own instance.

✅ Isomorphic: Comptabile with Node.js, Deno and modern browsers.

✅ Type Safe: Strongly typed and written in TypeScript.

✅ Scalable: Easily scales from a few calls to complex API networks with thousands of APIs.

✅ Tested: Battle tested in large projects, fully covered by unit tests.

✅ Maintained: Since 2021 publicly through Github.

✔️ Features

• 100% Performance-Oriented: Optimized for speed and efficiency, ensuring fast and reliable API interactions.
• Smart Retry Mechanism: Features exponential backoff for intelligent error handling and retry mechanisms.
• Automatic Request Deduplication: Set the time during which requests are deduplicated (treated as same request).
• Smart Cache Management: Dynamically manage cache with configurable expiration, custom keys, and selective invalidation.
• Dynamic URLs Support: Easily manage routes with dynamic parameters, such as /user/:userId.
• Error Handling: Flexible error management at both global and individual request levels.
• Automatic Request Cancellation: Utilizes AbortController to cancel previous requests automatically.
• Timeouts: Set timeouts globally or per request to prevent hanging operations.
• Multiple Fetching Strategies: Handle failed requests with various strategies - promise rejection, silent hang, soft fail, or default response.
• Multiple Requests Chaining: Easily chain multiple requests using promises for complex API interactions.
• Native fetch() Support: Utilizes the built-in fetch() API, providing a modern and native solution for making HTTP requests.
• Customizable: Fully compatible with a wide range of HTTP request configuration options, allowing for flexible and detailed request customization.
• Lightweight: Minimal footprint, only a few KBs when gzipped, ensuring quick load times.
• Framework Independent: Pure JavaScript solution, compatible with any framework or library.
• Cross-Framework compatible: Makes it easy to integration with Frameworks and Libraries, both Client Side and Server Side.
• Browser and Node.js 18+ Compatible: Works flawlessly in both modern browsers and Node.js environments.
• Fully TypeScript Compatible: Enjoy full TypeScript support for better development experience and type safety.
• Custom Interceptors: Includes onRequest, onResponse, and onError interceptors for flexible request and response handling.

Please open an issue for future requests.

Install

Using NPM:

npm install fetchff

Using Pnpm:

pnpm install fetchff

Using Yarn:

yarn add fetchff

✔️ API

Standalone usage

fetchf()

It is a functional wrapper for fetch(). It seamlessly enhances it with additional settings like the retry mechanism and error handling improvements. The fetchf() can be used directly as a function, simplifying the usage and making it easier to integrate with functional programming styles. The fetchf() makes requests independently of createApiFetcher() settings.

Example

import { fetchf } from 'fetchff';

const { data, error } = await fetchf('/api/user-details', {
  timeout: 5000,
  cancellable: true,
  retry: { retries: 3, delay: 2000 },
  // Specify some other settings here... The fetch() settings work as well...
});

Some of challenges with Native Fetch that fetchff solves:

• Error Status Handling: Fetch does not throw errors for HTTP error statuses, making it difficult to distinguish between successful and failed requests based on status codes alone.
• Error Visibility: Error responses with status codes like 404 or 500 are not automatically propagated as exceptions, which can lead to inconsistent error handling.
• No Built-in Retry Mechanism: Native fetch() lacks built-in support for retrying requests. Developers need to implement custom retry logic to handle transient errors or intermittent failures, which can be cumbersome and error-prone.
• Network Errors Handling: Native fetch() only rejects the Promise for network errors or failure to reach the server. Issues such as timeout errors or server unavailability do not trigger rejection by default, which can complicate error management.
• Limited Error Information: The error information provided by native fetch() is minimal, often leaving out details such as the request headers, status codes, or response bodies. This can make debugging more difficult, as there's limited visibility into what went wrong.
• Lack of Interceptors: Native fetch() does not provide a built-in mechanism for intercepting requests or responses. Developers need to manually manage request and response processing, which can lead to repetitive code and less maintainable solutions.
• No Built-in Caching: Native fetch() does not natively support caching of requests and responses. Implementing caching strategies requires additional code and management, potentially leading to inconsistencies and performance issues.

To address these challenges, the fetchf() provides several enhancements:

1. Consistent Error Handling:

   • In JavaScript, the native fetch() function does not reject the Promise for HTTP error statuses such as 404 (Not Found) or 500 (Internal Server Error). Instead, fetch() resolves the Promise with a Response object, where the ok property indicates the success of the request. If the request encounters a network error or fails due to other issues (e.g., server downtime), fetch() will reject the Promise.
   • This approach aligns error handling with common practices and makes it easier to manage errors consistently.

2. Enhanced Retry Mechanism:

   • Retry Configuration: You can configure the number of retries, delay between retries, and exponential backoff for failed requests. This helps to handle transient errors effectively.
   • Custom Retry Logic: The shouldRetry asynchronous function allows for custom retry logic based on the error and attempt count, providing flexibility to handle different types of failures.
   • Retry Conditions: Errors are only retried based on configurable retry conditions, such as specific HTTP status codes or error types.

3. Improved Error Visibility:

   • Error Wrapping: The createApiFetcher() and fetchf() wrap errors in a custom RequestError class, which provides detailed information about the request and response, similarly to what Axios does. This makes debugging easier and improves visibility into what went wrong.

4. Extended settings:

   • Check Settings table for more information about all settings.

Multiple API Endpoints

createApiFetcher()

It is a powerful factory function for creating API fetchers with advanced features. It provides a convenient way to configure and manage multiple API endpoints using a declarative approach. This function offers integration with retry mechanisms, error handling improvements, and all the other settings. Unlike traditional methods, createApiFetcher() allows you to set up and use API endpoints efficiently with minimal boilerplate code.

Example

import { createApiFetcher } from 'fetchff';

// Create some endpoints declaratively
const api = createApiFetcher({
  baseURL: 'https://example.com/api',
  endpoints: {
    getUser: {
      url: '/user-details/:id/',
      method: 'GET',
      // Each endpoint accepts all settings declaratively
      retry: { retries: 3, delay: 2000 },
      timeout: 5000,
      cancellable: true,
    },
    // Define more endpoints as needed
  },
  // You can set all settings globally
  strategy: 'softFail', // no try/catch required in case of errors
});

// Make a GET request to http://example.com/api/user-details/2/?rating[]=1&rating[]=2
const { data, error } = await api.getUser({
  params: { rating: [1, 2] }, //  Passed arrays, objects etc. will be parsed automatically
  urlPathParams: { id: 2 }, // Replace :id with 2 in the URL
});

Multiple API Specific Settings

All the Request Settings can be directly used in the function as global settings for all endpoints. They can be also used within the endpoints property (on per-endpoint basis). The exposed endpoints property is as follows:

• endpoints: Type: EndpointsConfig<EndpointsMethods> List of your endpoints. Each endpoint is an object that accepts all the Request Settings (see the Basic Settings below). The endpoints are required to be specified.

How It Works

The createApiFetcher() automatically creates and returns API methods based on the endpoints object provided. It also exposes some extra methods and properties that are useful to handle global config, dynamically add and remove endpoints etc.

api.yourEndpoint(requestConfig)

Where yourEndpoint is the name of your endpoint, the key from endpoints object passed to the createApiFetcher().

requestConfig (optional) object - To have more granular control over specific endpoints you can pass Request Config for particular endpoint. Check Basic Settings below for more information.

Returns: Response Object (see below).

api.request(endpointNameOrUrl, requestConfig)

The api.request() helper function is a versatile method provided for making API requests with customizable configurations. It allows you to perform HTTP requests to any endpoint defined in your API setup and provides a straightforward way to handle queries, path parameters, and request configurations dynamically.

Example

import { createApiFetcher } from 'fetchff';

const api = createApiFetcher({
  apiUrl: 'https://example.com/api',
  endpoints: {
    updateUser: {
      url: '/update-user/:id',
      method: 'POST',
    },
    // Define more endpoints as needed
  },
});

// Using api.request to make a POST request
const { data, error } = await api.request('updateUser', {
  body: {
    name: 'John Doe', // Data Payload
  },
  urlPathParams: {
    id: '123', // URL Path Param :id will be replaced with 123
  },
});

// Using api.request to make a GET request to an external API
const { data, error } = await api.request('https://example.com/api/user', {
  params: {
    name: 'John Smith', // Query Params
  },
});

api.config

You can access api.config property directly, so to modify global headers, and other settings on fly. Please mind it is a property, not a function.

api.endpoints

You can access api.endpoints property directly, so to modify endpoints list. It can be useful if you want to append or remove global endpoints. Please mind it is a property, not a function.

api.getInstance()

If you initialize API handler with your custom fetcher, then this function will return the instance is created using fetcher.create() function. Your fetcher can include anything. It will be triggering fetcher.request() instead of native fetch() that is available by default. It gives you ultimate flexibility on how you want your requests to be made.

🛠️ Plugin API Architecture

⚙️ Basic Settings

You can pass the settings:

• globally for all requests when calling createApiFetcher()
• per-endpoint basis defined under endpoints in global config when calling createApiFetcher()
• per-request basis when calling fetchf() (second argument of the function) or in the api.yourEndpoint() (third argument)

You can also use all native fetch() settings.

| | Type | Default | Description | | -------------------------- | ------------------------------------------------------------------------------------------------------ | ------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | baseURL(alias: apiUrl) | string | | Your API base url. | | url | string | | URL path e.g. /user-details/get | | method | string | GET | Default request method e.g. GET, POST, DELETE, PUT etc. All methods are supported. | | params | objectURLSearchParamsNameValuePair[] | {} | Query Parameters - a key-value pairs added to the URL to send extra information with a request. If you pass an object, it will be automatically converted. It works with nested objects, arrays and custom data structures similarly to what jQuery used to do in the past. If you use createApiFetcher() then it is the first argument of your api.yourEndpoint() function. You can still pass configuration in 3rd argument if want to.You can pass key-value pairs where the values can be strings, numbers, or arrays. For example, if you pass { foo: [1, 2] }, it will be automatically serialized into foo[]=1&foo[]=2 in the URL. | | body(alias: data) | objectstringFormDataURLSearchParamsBlobArrayBufferReadableStream | {} | The body is the data sent with the request, such as JSON, text, or form data, included in the request payload for POST, PUT, or PATCH requests. | | urlPathParams | object | {} | It lets you dynamically replace segments of your URL with specific values in a clear and declarative manner. This feature is especially handy for constructing URLs with variable components or identifiers.For example, suppose you need to update user details and have a URL template like /user-details/update/:userId. With urlPathParams, you can replace :userId with a real user ID, such as 123, resulting in the URL /user-details/update/123. | | flattenResponse | boolean | false | When set to true, this option flattens the nested response data. This means you can access the data directly without having to use response.data.data. It works only if the response structure includes a single data property. | | defaultResponse | any | null | Default response when there is no data or when endpoint fails depending on the chosen strategy | | withCredentials | boolean | false | Indicates whether credentials (such as cookies) should be included with the request. | | timeout | number | 30000 | You can set a request timeout for all requests or particular in milliseconds. | | dedupeTime | number | 1000 | Time window, in milliseconds, during which identical requests are deduplicated (treated as single request). | | logger | Logger | null | You can additionally specify logger object with your custom logger to automatically log the errors to the console. It should contain at least error and warn functions. | | fetcher | FetcherInstance | | A custom adapter (an instance / object) that exposes create() function so to create instance of API Fetcher. The create() should return request() function that would be used when making the requests. The native fetch() is used if the fetcher is not provided. |

🏷️ Headers

fetchff provides robust support for handling HTTP headers in your requests. You can configure and manipulate headers at both global and per-request levels. Here’s a detailed overview of how to work with headers using fetchff.

Note: Header keys are case-sensitive when specified in request objects. Ensure that the keys are provided in the correct case to avoid issues with header handling.

Setting Headers Globally

You can set default headers that will be included in all requests made with a specific createApiFetcher instance. This is useful for setting common headers like authentication tokens or content types.

Example: Setting Headers Globally

import { createApiFetcher } from 'fetchff';

const api = createApiFetcher({
  baseURL: 'https://api.example.com/',
  headers: {
    'Content-Type': 'application/json',
    Authorization: 'Bearer YOUR_TOKEN',
  },
  // other configurations
});

Setting Per-Request Headers

In addition to global default headers, you can also specify headers on a per-request basis. This allows you to override global headers or set specific headers for individual requests.

Example: Setting Per-Request Headers

import { fetchf } from 'fetchff';

// Example of making a GET request with custom headers
const { data } = await fetchf('https://api.example.com/endpoint', {
  headers: {
    Authorization: 'Bearer YOUR_ACCESS_TOKEN',
    'Custom-Header': 'CustomValue',
  },
});

Default Headers

The fetchff plugin automatically injects a set of default headers into every request. These default headers help ensure that requests are consistent and include necessary information for the server to process them correctly.

• Content-Type: application/json;charset=utf-8 Specifies that the request body contains JSON data and sets the character encoding to UTF-8.

• Accept: application/json, text/plain, */* Indicates the media types that the client is willing to receive from the server. This includes JSON, plain text, and any other types.

• Accept-Encoding: gzip, deflate, br Specifies the content encoding that the client can understand, including gzip, deflate, and Brotli compression.

🌀 Interceptors

Example

const { data } = await fetchf('https://api.example.com/', {
  onRequest(config) {
    // Add a custom header before sending the request
    config.headers['Authorization'] = 'Bearer your-token';
  },
  onResponse(response) {
    // Log the response status
    console.log(`Response Status: ${response.status}`);
  },
  onError(error, config) {
    // Handle errors and log the request config
    console.error('Request failed:', error);
    console.error('Request config:', config);
  },
});

Configuration

The following options are available for configuring interceptors in the RequestHandler:

• onRequest:
  Type: RequestInterceptor | RequestInterceptor[]
  A function or an array of functions that are invoked before sending a request. Each function receives the request configuration object as its argument, allowing you to modify request parameters, headers, or other settings.
  Default: (config) => config (no modification).

• onResponse:
  Type: ResponseInterceptor | ResponseInterceptor[]
  A function or an array of functions that are invoked when a response is received. Each function receives the full response object, enabling you to process the response, handle status codes, or parse data as needed.
  Default: (response) => response (no modification).

• onError:
  Type: ErrorInterceptor | ErrorInterceptor[]
  A function or an array of functions that handle errors when a request fails. Each function receives the error and request configuration as arguments, allowing you to implement custom error handling logic or logging.
  Default: (error) => error (no modification).

How It Works

1. Request Interception:
   Before a request is sent, the onRequest interceptors are invoked. These interceptors can modify the request configuration, such as adding headers or changing request parameters.

2. Response Interception:
   Once a response is received, the onResponse interceptors are called. These interceptors allow you to handle the response data, process status codes, or transform the response before it is returned to the caller.

3. Error Interception:
   If a request fails and an error occurs, the onError interceptors are triggered. These interceptors provide a way to handle errors, such as logging or retrying requests, based on the error and the request configuration.

4. Custom Handling:
   Each interceptor function provides a flexible way to customize request and response behavior. You can use these functions to integrate with other systems, handle specific cases, or modify requests and responses as needed.

🔍 Error Handling

Example

Here's an example of how to configure error handling:

const { data } = await fetchf('https://api.example.com/', {
  strategy: 'reject', // Use 'reject' strategy for error handling (default)
});

Configuration

The strategy option can be configured with the following values: Default: reject.

• reject:
  Promises are rejected, and global error handling is triggered. You must use try/catch blocks to handle errors.

• softFail:
  Returns a response object with additional properties such as data, error, config, request, and headers when an error occurs. This approach avoids throwing errors, allowing you to handle error information directly within the response object without the need for try/catch blocks.

• defaultResponse:
  Returns a default response specified in case of an error. The promise will not be rejected. This can be used in conjunction with flattenResponse and defaultResponse: {} to provide sensible defaults.

• silent:
  Hangs the promise silently on error, useful for fire-and-forget requests without the need for try/catch. In case of an error, the promise will never be resolved or rejected, and any code after will never be executed. This strategy is useful for dispatching requests within asynchronous wrapper functions that do not need to be awaited. It prevents excessive usage of try/catch or additional response data checks everywhere. It can be used in combination with onError to handle errors separately.

How It Works

1. Reject Strategy:
   When using the reject strategy, if an error occurs, the promise is rejected, and global error handling logic is triggered. You must use try/catch to handle these errors.

2. Soft Fail Strategy:
   With softFail, the response object includes additional properties that provide details about the error without rejecting the promise. This allows you to handle error information directly within the response.

3. Default Response Strategy:
   The defaultResponse strategy returns a predefined default response when an error occurs. This approach prevents the promise from being rejected, allowing for default values to be used in place of error data.

4. Silent Strategy:
   The silent strategy results in the promise hanging silently on error. The promise will not be resolved or rejected, and any subsequent code will not execute. This is useful for fire-and-forget requests and can be combined with onError for separate error handling.

5. Custom Error Handling:
   Depending on the strategy chosen, you can tailor how errors are managed, either by handling them directly within response objects, using default responses, or managing them silently.

🗄️ Smart Cache Management

Example

const { data } = await fetchf('https://api.example.com/', {
  cacheTime: 300, // Cache is valid for 5 minutes
  cacheKey: (config) => `${config.url}-${config.method}`, // Custom cache key based on URL and method
  cacheBuster: (config) => config.method === 'POST', // Bust cache for POST requests
  skipCache: (response, config) => response.status !== 200, // Skip caching on non-200 responses
});

Configuration

The caching system can be fine-tuned using the following options when configuring the:

• cacheTime:
  Type: number
  Specifies the duration, in seconds, for which a cache entry is considered "fresh." Once this time has passed, the entry is considered stale and may be refreshed with a new request.
  Default: 0 (no caching).

• cacheKey:
  Type: CacheKeyFunction
  A function used to generate a custom cache key for the request. If not provided, a default key is created by hashing various parts of the request, including Method, URL, query parameters, and headers.
  Default: Auto-generated based on request properties.

• cacheBuster:
  Type: CacheBusterFunction
  A function that allows you to invalidate or refresh the cache under certain conditions, such as specific request methods or response properties. This is useful for ensuring that certain requests (e.g., POST) bypass the cache.
  Default: (config) => false (no cache busting).

• skipCache:
  Type: CacheSkipFunction
  A function that determines whether caching should be skipped based on the response. This allows for fine-grained control over whether certain responses are cached or not, such as skipping non-200 responses.
  Default: (response, config) => false (no skipping).

How It Works

1. Request and Cache Check:
   When a request is made, the cache is first checked for an existing entry. If a valid cache entry is found and is still "fresh" (based on cacheTime), the cached response is returned immediately. Note that when the native fetch() setting called cache is set to reload the request will automatically skip the internal cache.

2. Cache Key:
   A cache key uniquely identifies each request. By default, the key is generated based on the URL and other relevant request options. Custom keys can be provided using the cacheKey function.

3. Cache Busting:
   If the cacheBuster function is defined, it determines whether to invalidate and refresh the cache for specific requests. This is useful for ensuring that certain requests, such as POST requests, always fetch new data.

4. Skipping Cache:
   The skipCache function provides flexibility in deciding whether to store a response in the cache. For example, you might skip caching responses that have a 4xx or 5xx status code.

5. Final Outcome:
   If no valid cache entry is found, or the cache is skipped or busted, the request proceeds to the network, and the response is cached based on the provided configuration.

✋ Automatic Request Cancellation

• Automatic retries for failed requests with configurable delay and exponential backoff.
• Optional flattening of response data for easier access, removing nested data fields.

You can choose to reject cancelled requests or return a default response instead through the defaultResponse setting.

Example

import { fetchf } from 'fetchff';

// Function to send the request
const sendRequest = () => {
  // In this example, the previous requests are automatically cancelled
  // You can also control "dedupeTime" setting in order to fire the requests more or less frequently
  fetchf('https://example.com/api/messages/update', {
    method: 'POST',
    cancellable: true,
    rejectCancelled: true,
  });
};

// Attach keydown event listener to the input element with id "message"
document.getElementById('message')?.addEventListener('keydown', sendRequest);

Configuration

• cancellable: Type: boolean Default: false If set to true, any ongoing previous requests to the same API endpoint will be automatically cancelled when a subsequent request is made before the first one completes. This is useful in scenarios where repeated requests are made to the same endpoint (e.g., search inputs) and only the latest response is needed, avoiding unnecessary requests to the backend.

• rejectCancelled: Type: boolean Default: false Works in conjunction with the cancellable option. If set to true, the promise of a cancelled request will be rejected. By default (false), when a request is cancelled, instead of rejecting the promise, a defaultResponse will be returned, allowing graceful handling of cancellation without errors.

📶 Polling Configuration

Example

Here's an example of how to configure polling:

const { data } = await fetchf('https://api.example.com/', {
  pollingInterval: 5000, // Poll every 5 seconds
  shouldStopPolling: (response, error, attempt) => {
    if (response && response.status === 200) {
      return true; // Stop polling if the response status is 200 (OK)
    }
    if (attempt >= 10) {
      return true; // Stop polling after 10 attempts
    }
    return false; // Continue polling otherwise
  },
});

Configuration

The following options are available for configuring polling in the RequestHandler:

• pollingInterval:
  Type: number
  Interval in milliseconds between polling attempts. If set to 0, polling is disabled. This allows you to control the frequency of requests when polling is enabled.
  Default: 0 (polling disabled).

• shouldStopPolling:
  Type: (response: any, error: any, attempt: number) => boolean
  A function to determine if polling should stop based on the response, error, or the current polling attempt number. Return true to stop polling, and false to continue polling. This allows for custom logic to decide when to stop polling based on the conditions of the response or error.
  Default: (response, error, attempt) => false (polling continues indefinitely unless manually stopped).

How It Works

1. Polling Interval:
   When pollingInterval is set to a non-zero value, polling begins after the initial request. The request is repeated at intervals defined by the pollingInterval setting.

2. Stopping Polling:
   The shouldStopPolling function is invoked after each polling attempt. If it returns true, polling will stop. Otherwise, polling will continue until the condition to stop is met, or polling is manually stopped.

3. Custom Logic:
   The shouldStopPolling function provides flexibility to implement custom logic based on the response, error, or the number of attempts. This makes it easy to stop polling when the desired outcome is reached or after a maximum number of attempts.

🔄 Retry Mechanism

Example

const { data } = await fetchf('https://api.example.com/', {
  retry: {
    retries: 3,
    delay: 100,
    maxDelay: 5000,
    resetTimeout: true,
    backoff: 1.5,
    retryOn: [500, 503],
    shouldRetry(error, attempt) {
      // Retry on specific errors or based on custom logic
      return attempt < 3; // Retry up to 3 times
    },
  },
});

Configuration

The retry mechanism is configured via the retry option when instantiating the RequestHandler. You can customize the following parameters:

• retries:
  Type: number
  Number of retry attempts to make after an initial failure.
  Default: 0 (no retries).

• delay:
  Type: number
  Initial delay (in milliseconds) before the first retry attempt. Subsequent retries use an exponentially increasing delay based on the backoff parameter.
  Default: 1000 (1 second).

• maxDelay:
  Type: number
  Maximum delay (in milliseconds) between retry attempts. The delay will not exceed this value, even if the exponential backoff would suggest a longer delay.
  Default: 30000 (30 seconds).

• backoff:
  Type: number
  Factor by which the delay is multiplied after each retry. For example, a backoff factor of 1.5 means each retry delay is 1.5 times the previous delay.
  Default: 1.5.

• retryOn:
  Type: number[]
  Array of HTTP status codes that should trigger a retry. By default, retries are triggered for the following status codes:

  • 408 - Request Timeout
  • 409 - Conflict
  • 425 - Too Early
  • 429 - Too Many Requests
  • 500 - Internal Server Error
  • 502 - Bad Gateway
  • 503 - Service Unavailable
  • 504 - Gateway Timeout

• shouldRetry:
  Type: RetryFunction
  Function that determines whether a retry should be attempted based on the error and the current attempt number. This function receives the error object and the attempt number as arguments.
  Default: Retry up to the number of specified attempts.

How It Works

1. Initial Request: When a request fails, the retry mechanism captures the failure and checks if it should retry based on the retryOn configuration and the result of the shouldRetry function.

2. Retry Attempts: If a retry is warranted:

   • The request is retried up to the specified number of attempts (retries).
   • Each retry waits for a delay before making the next attempt. The delay starts at the initial delay value and increases exponentially based on the backoff factor, but will not exceed the maxDelay.

3. Logging: During retries, the mechanism logs warnings indicating the retry attempts and the delay before the next attempt, which helps in debugging and understanding the retry behavior.

4. Final Outcome: If all retry attempts fail, the request will throw an error, and the final failure is processed according to the configured error handling logic.

🧩 Response Data Transformation

The fetchff plugin automatically handles response data transformation for any instance of Response returned by the fetch() (or a custom fetcher) based on the Content-Type header, ensuring that data is parsed correctly according to its format.

How It Works

• JSON (application/json): Parses the response as JSON.
• Form Data (multipart/form-data): Parses the response as FormData.
• Binary Data (application/octet-stream): Parses the response as a Blob.
• URL-encoded Form Data (application/x-www-form-urlencoded): Parses the response as FormData.
• Text (text/*): Parses the response as plain text.

If the Content-Type header is missing or not recognized, the plugin defaults to attempting JSON parsing. If that fails, it will try to parse the response as text.

This approach ensures that the fetchff plugin can handle a variety of response formats, providing a flexible and reliable method for processing data from API requests.

onResponse Interceptor

You can use the onResponse interceptor to customize how the response is handled before it reaches your application. This interceptor gives you access to the raw Response object, allowing you to transform the data or modify the response behavior based on your needs.

📄 Response Object

Structure of the Response Object

• data:

  • Type: ResponseData (or your custom type passed through generic)

  • Contains the actual data returned from the API request, null or value of defaultResponse setting, if nothing is found.

• error:

  • Type: ResponseErr

  • An object with details about any error that occurred or null otherwise.

  • name: The name of the error (e.g., 'ResponseError').

  • message: A descriptive message about the error.

  • status: The HTTP status code of the response (e.g., 404, 500).

  • statusText: The HTTP status text of the response (e.g., 'Not Found', 'Internal Server Error').

  • request: Details about the HTTP request that was sent (e.g., URL, method, headers).

  • config: The configuration object used for the request, including URL, method, headers, and query parameters.

  • response: The full response object received from the server, including all headers and body.

• config:

  • Type: RequestConfig
  • The configuration object with all settings used for the request, including URL, method, headers, and query parameters.

• status:

  • Type: number
  • The HTTP status code of the response (e.g., 404, 500).

• statusText:

  • Type: string
  • The HTTP status text of the response (e.g., 'Not Found', 'Internal Server Error').

• request:

  • Type: RequestConfig
  • An alias for config.

• headers:

  • Type: HeadersObject
  • The response headers returned by the server, such as content type and caching information returned as simple key-value object.

The whole response of the native fetch() is attached as well.

📦 Typings

The fetchff package provides comprehensive TypeScript typings to enhance development experience and ensure type safety. Below are details on the available, exportable types for both createApiFetcher() and fetchf().

Generic Typings

The fetchff package includes several generic types to handle various aspects of API requests and responses:

• QueryParams<ParamsType>: Represents query parameters for requests. Can be an object, URLSearchParams, an array of name-value pairs, or null.
• BodyPayload<PayloadType>: Represents the request body. Can be BodyInit, an object, an array, a string, or null.
• UrlPathParams<UrlParamsType>: Represents URL path parameters. Can be an object or null.
• DefaultResponse: Default response for all requests. Default is: any.

Typings for createApiFetcher()

The createApiFetcher<EndpointsMethods, EndpointsConfiguration>() function provides a robust set of types to define and manage API interactions.

The key types are:

• EndpointsMethods: Represents the list of API endpoints with their respective settings. It is your own interface that you can pass to this generic. It will be cross-checked against the endpoints object in your createApiFetcher() configuration.Each endpoint can be configured with its own specific settings such as Response Payload, Query Parameters and URL Path Parameters.
• Endpoint<ResponseData = DefaultResponse, QueryParams = DefaultParams, PathParams = DefaultUrlParams, RequestBody = DefaultPayload>: Represents an API endpoint function, allowing to be defined with optional query parameters, URL path parameters, request configuration (settings), and request body (data).
• EndpointsSettings: Configuration for API endpoints, including query parameters, URL path parameters, and additional request configurations. Default is typeof endpoints.
• RequestInterceptor: Function to modify request configurations before they are sent.
• ResponseInterceptor: Function to process responses before they are handled by the application.
• ErrorInterceptor: Function to handle errors when a request fails.
• CreatedCustomFetcherInstance: Represents the custom fetcher instance created by its create() function.

For a full list of types and detailed definitions, refer to the api-handler.ts file.

Typings for fetchf()

The fetchf() function includes types that help configure and manage network requests effectively:

• RequestHandlerConfig: Main configuration options for the fetchf() function, including request settings, interceptors, and retry configurations.
• RetryConfig: Configuration options for retry mechanisms, including the number of retries, delay between retries, and backoff strategies.
• CacheConfig: Configuration options for caching, including cache time, custom cache keys, and cache invalidation rules.
• PollingConfig: Configuration options for polling, including polling intervals and conditions to stop polling.
• ErrorStrategy: Defines strategies for handling errors, such as rejection, soft fail, default response, and silent modes.

For a complete list of types and their definitions, refer to the request-handler.ts file.

Benefits of Using Typings

• Type Safety: Ensures that configurations and requests adhere to expected formats, reducing runtime errors and improving reliability.
• Autocompletion: Provides better support for autocompletion in editors, making development faster and more intuitive.
• Documentation: Helps in understanding available options and their expected values, improving code clarity and maintainability.

Comparison with another libraries

| Feature | fetchff | ofetch | wretch | axios | native fetch() | | -------------------------------------------------- | ----------- | ------------ | ------------ | ------------ | -------------- | | Unified API Client | ✅ | -- | -- | -- | -- | | Smart Request Cache | ✅ | -- | -- | -- | -- | | Automatic Request Deduplication | ✅ | -- | -- | -- | -- | | Custom Fetching Adapter | ✅ | -- | -- | -- | -- | | Built-in Error Handling | ✅ | -- | ✅ | -- | -- | | Customizable Error Handling | ✅ | -- | ✅ | ✅ | -- | | Retries with exponential backoff | ✅ | -- | -- | -- | -- | | Advanced Query Params handling | ✅ | -- | -- | -- | -- | | Custom Retry logic | ✅ | ✅ | ✅ | -- | -- | | Easy Timeouts | ✅ | ✅ | ✅ | ✅ | -- | | Polling Functionality | ✅ | -- | -- | -- | -- | | Easy Cancellation of stale (previous) requests | ✅ | -- | -- | -- | -- | | Default Responses | ✅ | -- | -- | -- | -- | | Custom adapters (fetchers) | ✅ | -- | -- | ✅ | -- | | Global Configuration | ✅ | -- | ✅ | ✅ | -- | | TypeScript Support | ✅ | ✅ | ✅ | ✅ | ✅ | | Built-in AbortController Support | ✅ | -- | -- | -- | -- | | Request Interceptors | ✅ | ✅ | ✅ | ✅ | -- | | Request and Response Transformation | ✅ | ✅ | ✅ | ✅ | -- | | Integration with Libraries | ✅ | ✅ | ✅ | ✅ | -- | | Request Queuing | ✅ | -- | -- | -- | -- | | Multiple Fetching Strategies | ✅ | -- | -- | -- | -- | | Dynamic URLs | ✅ | -- | ✅ | -- | -- | | Automatic Retry on Failure | ✅ | ✅ | -- | ✅ | -- | | Server-Side Rendering (SSR) Support | ✅ | ✅ | -- | -- | -- | | Minimal Installation Size | 🟢 (3.3 KB) | 🟡 (6.41 KB) | 🟢 (2.21 KB) | 🔴 (13.7 KB) | 🟢 (0 KB) |

✏️ Examples

Click to expand particular examples below. You can also check examples.ts for more examples of usage.

All Settings

Here’s an example of configuring and using the createApiFetcher() with all available settings.

const api = createApiFetcher({
  baseURL: 'https://api.example.com/',
  endpoints: {
    getBooks: {
      url: 'books/all',
      method: 'get',
      cancellable: true,
      // All the global settings can be specified on per-endpoint basis as well
    },
  },
  strategy: 'reject', // Error handling strategy.
  cancellable: false, // If true, cancels previous requests to same endpoint.
  rejectCancelled: false, // Reject promise for cancelled requests.
  flattenResponse: false, // If true, flatten nested response data.
  defaultResponse: null, // Default response when there is no data or endpoint fails.
  withCredentials: true, // Pass cookies to all requests.
  timeout: 30000, // Request timeout in milliseconds.
  dedupeTime: 1000, // Time window, in milliseconds, during which identical requests are deduplicated (treated as single request).
  pollingInterval: 5000, // Interval in milliseconds between polling attempts. Setting 0 disables polling.
  shouldStopPolling: (response, error, attempt) => false, // Function to determine if polling should stop based on the response. Returns true to stop polling, false to continue.
  method: 'get', // Default request method.
  params: {}, // Default params added to all requests.
  data: {}, // Alias for 'body'. Default data passed to POST, PUT, DELETE and PATCH requests.
  cacheTime: 300, // Cache is valid for 5 minutes
  cacheKey: (config) => `${config.url}-${config.method}`, // Custom cache key based on URL and method
  cacheBuster: (config) => config.method === 'POST', // Bust cache for POST requests
  skipCache: (response, config) => response.status !== 200, // Skip caching on non-200 responses
  onError(error) {
    // Interceptor on error
    console.error('Request failed', error);
  },
  async onRequest(config) {
    // Interceptor on each request
    console.error('Fired on each request', config);
  },
  async onResponse(response) {
    // Interceptor on each response
    console.error('Fired on each response', response);
  },
  logger: {
    // Custom logger for logging errors.
    error(...args) {
      console.log('My custom error log', ...args);
    },
    warn(...args) {
      console.log('My custom warning log', ...args);
    },
  },
  retry: {
    retries: 3, // Number of retries on failure.
    delay: 1000, // Initial delay between retries in milliseconds.
    backoff: 1.5, // Backoff factor for retry delay.
    maxDelay: 30000, // Maximum delay between retries in milliseconds.
    resetTimeout: true, // Reset the timeout when retrying requests.
    retryOn: [408, 409, 425, 429, 500, 502, 503, 504], // HTTP status codes to retry on.
    shouldRetry: async (error, attempts) => {
      // Custom retry logic.
      return (
        attempts < 3 &&
        [408, 500, 502, 503, 504].includes(error.response.status)
      );
    },
  },
});

try {
  // The same API config as used above, except the "endpoints" and "fetcher" and fetcher could be used as 3rd argument of the api.getBooks()
  const { data } = await api.getBooks();
  console.log('Request succeeded:', data);
} catch (error) {
  console.error('Request ultimately failed:', error);
}

Examples Using createApiFetcher()

All examples below are with usage of createApiFetcher(). You can also use fetchf() independently.

Multiple APIs Handler from different API sources

import { createApiFetcher } from 'fetchff';

// Create fetcher instance
const api = createApiFetcher({
  baseURL: 'https://example.com/api/v1',
  endpoints: {
    sendMessage: {
      method: 'post',
      url: '/send-message/:postId',
    },
    getMessage: {
      url: '/get-message/',
      // Change baseURL to external for this endpoint onyl
      baseURL: 'https://externalprovider.com/api/v2',
    },
  },
});

// Make a wrapper function and call your API
async function sendAndGetMessage() {
  await api.sendMessage({
    body: { message: 'Text' },
    urlPathParams: { postId: 1 },
  });

  const { data } = await api.getMessage({
    params: { postId: 1 },
  });
}

// Invoke your wrapper function
sendAndGetMessage();

Using with Full TypeScript Support

The library includes all necessary TypeScript definitions bringing full TypeScript support to your API Handler. The package ships interfaces with responsible defaults making it easier to add new endpoints.

// books.d.ts
interface Book {
  id: number;
  title: string;
  rating: number;
}

interface Books {
  books: Book[];
  totalResults: number;
}

interface BookQueryParams {
  newBook: boolean;
}

interface BookPathParams {
  bookId?: number;
}

// api.ts
import type { Endpoint } from 'fetchff';
import { createApiFetcher } from 'fetchff';

const endpoints = {
  fetchBooks: {
    url: 'books',
  },
  fetchBook: {
    url: 'books/:bookId',
  },
};

// No need to specify all endpoints types. For example, the "fetchBooks" is inferred automatically.
interface EndpointsList {
  fetchBook: Endpoint<Book, BookQueryParams, BookPathParams>;
}

type EndpointsConfiguration = typeof endpoints;

const api = createApiFetcher<EndpointsList, EndpointsConfiguration>({
  apiUrl: 'https://example.com/api/',
  endpoints,
});

const book = await api.fetchBook({
  params: { newBook: true },
  urlPathParams: { bookId: 1 },
});

// Will return an error since "rating" does not exist in "BookQueryParams"
const anotherBook = await api.fetchBook({ params: { rating: 5 } });

// You can also pass generic type directly to the request
const books = await api.fetchBooks<Books>();

Using with TypeScript and Custom Headers

import { createApiFetcher } from 'fetchff';

const endpoints = {
  getPosts: {
    url: '/posts/:subject',
  },

  getUser: {
    // Generally there is no need to specify method: 'get' for GET requests as it is default one. It can be adjusted using global "method" setting
    method: 'get',
    url: '/user-details',
  },

  updateUserDetails: {
    method: 'post',
    url: '/user-details/update/:userId',
    strategy: 'defaultResponse',
  },
};

interface EndpointsList {
  getPosts: Endpoint<PostsResponse, PostsQueryParams, PostsPathParams>;
}

type EndpointsConfiguration = typeof endpoints;

const api = createApiFetcher<EndpointsList, EndpointsConfiguration>({
  apiUrl: 'https://example.com/api',
  endpoints,
  onError(error) {
    console.log('Request failed', error);
  },
  headers: {
    'my-auth-key': 'example-auth-key-32rjjfa',
  },
});

// Fetch user data - "data" will return data directly
// GET to: http://example.com/api/user-details?userId=1&ratings[]=1&ratings[]=2
const { data } = await api.getUser({ params: { userId: 1, ratings: [1, 2] } });

// Fetch posts - "data" will return data directly
// GET to: http://example.com/api/posts/myTestSubject?additionalInfo=something
const { data } = await api.getPosts({
  params: { additionalInfo: 'something' },
  urlPathParams: { subject: 'test' },
});

// Send POST request to update userId "1"
await api.updateUserDetails({
  body: { name: 'Mark' },
  urlPathParams: { userId: 1 },
});

// Send POST request to update array of user ratings for userId "1"
await api.updateUserDetails({
  body: { name: 'Mark', ratings: [1, 2] },
  urlPathParams: { userId: 1 },
});

In the example above we fetch data from an API for user with an ID of 1. We also make a GET request to fetch some posts, update user's name to Mark. If you want to use more strict typings, please check TypeScript Usage section below.

Custom Fetcher

import { createApiFetcher, RequestConfig, FetchResponse } from 'fetchff';

// Define the custom fetcher object
const customFetcher = {
  create() {
    // Create instance here. It will be called at the beginning of every request.
    return {
      // This function will be called whenever a request is being fired.
      request: async (config: RequestConfig): Promise<FetchResponse> => {
        // Implement your custom fetch logic here
        const response = await fetch(config.url, config);
        // Optionally, process or transform the response
        return response;
      },
    };
  },
};

// Create the API fetcher with the custom fetcher
const api = createA