https://github.com/mattccc/fetchff
Fetchff is a lightweight, powerful and flexible HTTP client library designed to simplify request handling.
https://github.com/mattccc/fetchff
ajax api api-handler axios-api-handler axios-instance axios-multi-api cache fetch fetchf fetchff http http-client http-request javascript nodejs promise promises request typescript typescript-support
Last synced: 4 months ago
JSON representation
Fetchff is a lightweight, powerful and flexible HTTP client library designed to simplify request handling.
- Host: GitHub
- URL: https://github.com/mattccc/fetchff
- Owner: MattCCC
- License: unlicense
- Created: 2021-02-12T11:28:59.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2025-06-08T10:47:50.000Z (12 months ago)
- Last Synced: 2025-06-08T11:28:50.410Z (12 months ago)
- Topics: ajax, api, api-handler, axios-api-handler, axios-instance, axios-multi-api, cache, fetch, fetchf, fetchff, http, http-client, http-request, javascript, nodejs, promise, promises, request, typescript, typescript-support
- Language: TypeScript
- Homepage: https://www.npmjs.com/package/fetchff
- Size: 2.41 MB
- Stars: 37
- Watchers: 2
- Forks: 3
- Open Issues: 8
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- Funding: .github/FUNDING.yml
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
- Security: SECURITY.md
Awesome Lists containing this project
README
Fast, lightweight (~5 KB gzipped) and reusable data fetching
"fetchff" stands for "fetch fast & flexibly"
[npm-url]: https://npmjs.org/package/fetchff
[npm-image]: https://img.shields.io/npm/v/fetchff.svg
[![NPM version][npm-image]][npm-url] [](https://github.com/MattCCC/fetchff) [](https://github.com/MattCCC/fetchff) [](http://npm-stat.com/charts.html?package=fetchff) [](https://bundlephobia.com/result?p=fetchff) [](https://security.snyk.io/package/npm/fetchff)
## Why?
This is a high level library to extend the functionality of native fetch() with everything necessary and no overhead, so to wrap and reuse common patterns and functionalities in a simple and declarative manner. It is designed to be used in high-throughput, high-performance applications.
Also, managing multitude of API connections in large applications 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.
Click to expand
**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.
- The `fetchff` plugin aligns error handling with common practices and makes it easier to manage errors consistently by rejecting erroneous status codes.
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 from `response.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 `ResponseError` class, which provides detailed information about the request and response. This makes debugging easier and improves visibility into what went wrong.
4. **Extended settings:**
- Check Settings table for more information about all settings.
## ✔️ Benefits
✅ **Lightweight:** Minimal code footprint of ~4KB gzipped for managing extensive APIs.
✅ **High-Performance**: Optimized for speed and efficiency, ensuring fast and reliable API interactions.
✅ **Secure:** Secure by default rather than "permissive by default", with built-in sanitization mechanisms.
✅ **Immutable:** Every request has its own instance.
✅ **Isomorphic:** Compatible 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.
✅ **Customizable:** Fully compatible with a wide range configuration options, allowing for flexible and detailed request customization.
✅ **Responsible Defaults:** All settings are opt-in.
✅ **Framework Independent**: Pure JavaScript solution, compatible with any framework or library, both client and server side.
✅ **Browser and Node.js 18+ Compatible:** Works flawlessly in both modern browsers and Node.js environments.
✅ **Maintained:** Since 2021 publicly through Github.
## ✔️ Features
- **Smart Retry Mechanism**: Features exponential backoff for intelligent error handling and retry mechanisms.
- **Request Deduplication**: Set the time during which requests are deduplicated (treated as same request).
- **Cache Management**: Dynamically manage cache with configurable expiration, custom keys, and selective invalidation.
- **Network Revalidation**: Automatically revalidate data on window focus and network reconnection for fresh data.
- **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.
- **Request Cancellation**: Utilizes `AbortController` to cancel previous requests automatically.
- **Adaptive Timeouts**: Smart timeout adjustment based on connection speed for optimal user experience.
- **Fetching Strategies**: Handle failed requests with various strategies - promise rejection, silent hang, soft fail, or default response.
- **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.
- **Custom Interceptors**: Includes `onRequest`, `onResponse`, and `onError` interceptors for flexible request and response handling.
## ✔️ Install
[](https://npmjs.org/package/fetchff)
Using NPM:
```bash
npm install fetchff
```
Using Pnpm:
```bash
pnpm install fetchff
```
Using Yarn:
```bash
yarn add fetchff
```
## ✔️ API
### Standalone usage
#### `fetchf(url, config)`
_Alias: `fetchff(url, config)`_
A simple function that wraps the native `fetch()` and adds extra features like retries and better error handling. Use `fetchf()` directly for quick, enhanced requests - no need to set up `createApiFetcher()`. It works independently and is easy to use in any codebase.
#### Example
```typescript
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...
});
```
### Global Configuration
#### `getDefaultConfig()`
Click to expand
Returns the current global default configuration used for all requests. This is useful for inspecting or debugging the effective global settings.
```typescript
import { getDefaultConfig } from 'fetchff';
// Retrieve the current global default config
const config = getDefaultConfig();
console.log('Current global fetchff config:', config);
```
#### `setDefaultConfig(customConfig)`
Click to expand
Allows you to globally override the default configuration for all requests. This is useful for setting application-wide defaults like timeouts, headers, or retry policies.
```typescript
import { setDefaultConfig } from 'fetchff';
// Set global defaults for all requests
setDefaultConfig({
timeout: 10000, // 10 seconds for all requests
headers: {
Authorization: 'Bearer your-token',
},
retry: {
retries: 2,
delay: 1500,
},
});
// All subsequent requests will use these defaults
const { data } = await fetchf('/api/data'); // Uses 10s timeout and retry config
```
### Instance with many API endpoints
#### `createApiFetcher(config)`
Click to expand
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
```typescript
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`
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
```typescript
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 to modify global headers and other settings on the fly. This is a property, not a function.
#### `api.endpoints`
You can access `api.endpoints` property directly to modify the endpoints list. This can be useful if you want to append or remove global endpoints. This is a property, not a function.
### Advanced Utilities
Click to expand
#### Cache Management
##### `mutate(key, newData, settings)`
Programmatically update cached data without making a network request. Useful for optimistic updates or reflecting changes from other operations.
**Parameters:**
- `key` (string): The cache key to update
- `newData` (any): The new data to store in cache
- `settings` (object, optional): Configuration options
- `revalidate` (boolean): Whether to trigger background revalidation after update
```typescript
import { mutate } from 'fetchff';
// Update cache for a specific cache key
await mutate('/api/users', newUserData);
// Update with options
await mutate('/api/users', updatedData, {
revalidate: true, // Trigger background revalidation
});
```
##### `getCache(key)`
Directly retrieve cached data for a specific cache key. Useful for reading the current cached response without triggering a network request.
**Parameters:**
- `key` (string): The cache key to retrieve (equivalent to `cacheKey` from request config or `config.cacheKey` from response object)
**Returns:** The cached response object, or `null` if not found
```typescript
import { getCache } from 'fetchff';
// Get cached data for a specific key assuming you set {cacheKey: ''/api/user-profile'} in config
const cachedResponse = getCache('/api/user-profile');
if (cachedResponse) {
console.log('Cached user profile:', cachedResponse.data);
}
```
##### `setCache(key, response, ttl, staleTime)`
Directly set cache data for a specific key. Unlike `mutate()`, this doesn't trigger revalidation by default. This is a low level function to directly set cache data based on particular key. If unsure, use the `mutate()` with `revalidate: false` instead.
**Parameters:**
- `key` (string): The cache key to set. It must match the cache key of the request.
- `response` (any): The full response object to store in cache
- `ttl` (number, optional): Time to live for the cache entry, in seconds. Determines how long the cached data remains valid before expiring. If not specified, the default `0` value will be used (discard cache immediately), if `-1` specified then the cache will be held until manually removed using `deleteCache(key)` function.
- `staleTime` (number, optional): Duration, in seconds, for which cached data is considered "fresh" before it becomes eligible for background revalidation. If not specified, the default stale time applies.
```typescript
import { setCache } from 'fetchff';
// Set cache data with custom ttl and staleTime
setCache('/api/user-profile', userData, 600, 60); // Cache for 10 minutes, fresh for 1 minute
// Set cache for specific endpoint infinitely
setCache('/api/user-settings', userSettings, -1);
```
##### `deleteCache(key)`
Remove cached data for a specific cache key. Useful for cache invalidation when you know data is stale.
**Parameters:**
- `key` (string): The cache key to delete
```typescript
import { deleteCache } from 'fetchff';
// Delete specific cache entry
deleteCache('/api/user-profile');
// Delete cache after user logout
const logout = () => {
deleteCache('/api/user/*'); // Delete all user-related cache
};
```
#### Revalidation Management
##### `revalidate(key, isStaleRevalidation)`
Manually trigger revalidation for a specific cache entry, forcing a fresh network request to update the cached data.
**Parameters:**
- `key` (string): The cache key to revalidate
- `isStaleRevalidation` (boolean, optional): Whether this is a background revalidation that doesn't mark as in-flight
```typescript
import { revalidate } from 'fetchff';
// Revalidate specific cache entry
await revalidate('/api/user-profile');
// Revalidate with custom cache key
await revalidate('custom-cache-key');
// Background revalidation (doesn't mark as in-flight)
await revalidate('/api/user-profile', true);
```
##### `revalidateAll(type, isStaleRevalidation)`
Trigger revalidation for all cache entries associated with a specific event type (focus or online).
**Parameters:**
- `type` (string): The revalidation event type ('focus' or 'online')
- `isStaleRevalidation` (boolean, optional): Whether this is a background revalidation
```typescript
import { revalidateAll } from 'fetchff';
// Manually trigger focus revalidation for all relevant entries
revalidateAll('focus');
// Manually trigger online revalidation for all relevant entries
revalidateAll('online');
```
##### `removeRevalidators(type)`
Clean up revalidation event listeners for a specific event type. Useful for preventing memory leaks when you no longer need automatic revalidation.
**Parameters:**
- `type` (string): The revalidation event type to remove ('focus' or 'online')
```typescript
import { removeRevalidators } from 'fetchff';
// Remove all focus revalidation listeners
removeRevalidators('focus');
// Remove all online revalidation listeners
removeRevalidators('online');
// Typically called during cleanup
// e.g., in React useEffect cleanup or when unmounting components
```
#### Pub/Sub System
##### `subscribe(key, callback)`
Subscribe to cache updates and data changes. Receive notifications when specific cache entries are updated.
**Parameters:**
- `key` (string): The cache key to subscribe to
- `callback` (function): Function called when cache is updated
- `response` (any): The full response object
**Returns:** Function to unsubscribe from updates
```typescript
import { subscribe } from 'fetchff';
// Subscribe to cache changes for a specific key
const unsubscribe = subscribe('/api/user-data', (response) => {
console.log('Cache updated with response:', response);
console.log('Response data:', response.data);
console.log('Response status:', response.status);
});
// Clean up subscription when no longer needed
unsubscribe();
```
#### Request Management
##### `abortRequest(key, error)`
Programmatically abort in-flight requests for a specific cache key or URL pattern.
**Parameters:**
- `key` (string): The cache key or URL pattern to abort
- `error` (Error, optional): Custom error to throw for aborted requests
```typescript
import { abortRequest } from 'fetchff';
// Abort specific request by cache key
abortRequest('/api/slow-operation');
// Useful for cleanup when component unmounts or route changes
const cleanup = () => {
abortRequest('/api/user-dashboard');
};
```
#### Network Detection
##### `isSlowConnection()`
Check if the user is on a slow network connection (2G/3G). Useful for adapting application behavior based on connection speed.
**Parameters:** None
**Returns:** Boolean indicating if connection is slow
```typescript
import { isSlowConnection } from 'fetchff';
// Check connection speed and adapt behavior
if (isSlowConnection()) {
console.log('User is on a slow connection');
// Reduce image quality, disable auto-refresh, etc.
}
// Use in conditional logic
const shouldAutoRefresh = !isSlowConnection();
const imageQuality = isSlowConnection() ? 'low' : 'high';
```
## 🛠️ Plugin API Architecture
Click to expand
## ⚙️ Basic Settings
Click to expand
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](https://developer.mozilla.org/en-US/docs/Web/API/fetch#parameters).
| | Type | Default | Description |
| -------------------------- | ------------------------------------------------------------------------------------------------------ | ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| baseURL
(alias: apiUrl) | `string` | `undefined` | Your API base url. |
| url | `string` | `undefined` | 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 | `object`
`URLSearchParams`
`NameValuePair[]` | `undefined` | 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) | `object`
`string`
`FormData`
`URLSearchParams`
`Blob`
`ArrayBuffer`
`ReadableStream` | `undefined` | 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` | `undefined` | 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. |
| select | `(data: any) => any` | `undefined` | Function to transform or select a subset of the response data before it is returned. Called with the raw response data and should return the transformed data. Useful for mapping, picking, or shaping the response. |
| 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. This equals to `credentials: "include"` in native `fetch()`. In Node.js, cookies are not managed automatically. Use a fetch polyfill or library that supports cookies if needed. |
| timeout | `number` | `30000` / `60000` | You can set a request timeout in milliseconds. **Default is adaptive**: 30 seconds (30000 ms) for normal connections, 60 seconds (60000 ms) on slow connections (2G/3G). The timeout option applies to each individual request attempt including retries and polling. `0` means that the timeout is disabled. |
| dedupeTime | `number` | `0` | Time window, in milliseconds, during which identical requests are deduplicated (treated as same request). If set to `0`, deduplication is disabled. |
| cacheTime | `number` | `undefined` | 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. Set to -1 for indefinite cache. By default no caching. |
| staleTime | `number` | `undefined` | Specifies the duration, in seconds, for which cached data is considered "fresh." During this period, cached data will be returned immediately, but a background revalidation (network request) will be triggered to update the cache. If set to `0`, background revalidation is disabled and revalidation is triggered on every access. |
| refetchOnFocus | `boolean` | `false` | When set to `true`, automatically revalidates (refetches) data when the browser window regains focus. **Note: This bypasses the cache and always makes a fresh network request** to ensure users see the most up-to-date data when they return to your application from another tab or window. Particularly useful for applications that display real-time or frequently changing data, but should be used judiciously to avoid unnecessary network traffic. |
| refetchOnReconnect | `boolean` | `false` | When set to `true`, automatically revalidates (refetches) data when the browser regains internet connectivity after being offline. **This uses background revalidation to silently update data** without showing loading states to users. Helps ensure your application displays fresh data after network interruptions. Works by listening to the browser's `online` event. |
| 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 | `CustomFetcher` | `undefined` | A custom fetcher async function. By default, the native `fetch()` is used. If you use your own fetcher, default response parsing e.g. `await response.json()` call will be skipped. Your fetcher should return response object / data directly. |
| parser | `(response: Response) => Promise` | `undefined` | A custom response parser function. When provided, it replaces the built-in content-type based parsing entirely. Receives the raw `Response` object and should return the parsed data. Useful for handling custom formats like XML, CSV, or proprietary data. Can be set globally or per-request. |
> 📋 **Additional Settings Available:**
> The table above shows the most commonly used settings. Many more advanced configuration options are available and documented in their respective sections below, including:
>
> - **🔄 Retry Mechanism** - `retries`, `delay`, `maxDelay`, `backoff`, `resetTimeout`, `retryOn`, `shouldRetry`
> - **📶 Polling Configuration** - `pollingInterval`, `pollingDelay`, `maxPollingAttempts`, `shouldStopPolling`
> - **🗄️ Cache Management** - `cacheKey`, `cacheBuster`, `skipCache`, `cacheErrors`
> - **✋ Request Cancellation** - `cancellable`, `rejectCancelled`
> - **🌀 Interceptors** - `onRequest`, `onResponse`, `onError`, `onRetry`
> - **🔍 Error Handling** - `strategy`
### Performance Implications of Settings
Understanding the performance impact of different settings helps you optimize for your specific use case:
#### **High-Performance Settings**
**Minimize Network Requests:**
```typescript
// Aggressive caching for static data
const staticConfig = {
cacheTime: 3600, // 1 hour cache
staleTime: 1800, // 30 minutes freshness
dedupeTime: 10000, // 10 seconds deduplication
};
// Result: 90%+ reduction in network requests
```
**Optimize for Mobile/Slow Connections:**
```typescript
const mobileOptimized = {
timeout: 60000, // Longer timeout for slow connections (auto-adaptive)
retry: {
retries: 5, // More retries for unreliable connections
delay: 2000, // Longer initial delay (auto-adaptive)
backoff: 2.0, // Aggressive backoff
},
cacheTime: 900, // Longer cache on mobile
};
```
#### **Memory vs Network Trade-offs**
**Memory-Efficient (Low Cache):**
```typescript
const memoryEfficient = {
cacheTime: 60, // Short cache (1 minute)
staleTime: undefined, // No stale-while-revalidate
dedupeTime: 1000, // Short deduplication
};
// Pros: Low memory usage
// Cons: More network requests, slower perceived performance
```
**Network-Efficient (High Cache):**
```typescript
const networkEfficient = {
cacheTime: 1800, // Long cache (30 minutes)
staleTime: 300, // 5 minutes stale-while-revalidate
dedupeTime: 5000, // Longer deduplication
};
// Pros: Fewer network requests, faster user experience
// Cons: Higher memory usage, potentially stale data
```
#### **Feature Performance Impact**
| Feature | Performance Impact | Best Use Case |
| -------------------------- | ----------------------------------- | ------------------------------------------- |
| **Caching** | ⬇️ 70-90% fewer requests | Static or semi-static data |
| **Deduplication** | ⬇️ 50-80% fewer concurrent requests | High-traffic applications |
| **Stale-while-revalidate** | ⬆️ 90% faster perceived loading | Dynamic data that tolerates brief staleness |
| **Request cancellation** | ⬇️ Reduced bandwidth waste | Search-as-you-type, rapid navigation |
| **Retry mechanism** | ⬆️ 95%+ success rate | Mission-critical operations |
| **Polling** | ⬆️ Real-time updates | Live data monitoring |
#### **Adaptive Performance by Connection**
FetchFF automatically adapts timeouts and retry delays based on connection speed:
```typescript
// Automatic adaptation (no configuration needed)
const adaptiveRequest = fetchf('/api/data');
// On fast connections (WiFi/4G):
// - timeout: 30 seconds
// - retry delay: 1 second → 1.5s → 2.25s...
// - max retry delay: 30 seconds
// On slow connections (2G/3G):
// - timeout: 60 seconds
// - retry delay: 2 seconds → 3s → 4.5s...
// - max retry delay: 60 seconds
```
#### **Performance Patterns**
**Progressive Loading (Best UX):**
```typescript
// Layer 1: Instant response with cache
const quickData = await fetchf('/api/summary', {
cacheTime: 300,
staleTime: 60,
});
// Layer 2: Background enhancement
fetchf('/api/detailed-data', {
strategy: 'silent',
cacheTime: 600,
onResponse(response) {
updateUIWithDetailedData(response.data);
},
});
```
**Bandwidth-Conscious Loading:**
```typescript
// Check connection before expensive operations
import { isSlowConnection } from 'fetchff';
const loadUserDashboard = async () => {
const isSlowConn = isSlowConnection();
// Essential data always loads
const userData = await fetchf('/api/user', {
cacheTime: isSlowConn ? 600 : 300, // Longer cache on slow connections
});
// Optional data only on fast connections
if (!isSlowConn) {
fetchf('/api/user/analytics', { strategy: 'silent' });
fetchf('/api/user/recommendations', { strategy: 'silent' });
}
};
```
#### **Performance Monitoring**
Track key metrics to optimize your settings:
```typescript
const performanceConfig = {
onRequest(config) {
console.time(`request-${config.url}`);
},
onResponse(response) {
console.timeEnd(`request-${response.config.url}`);
// Track cache hit rate
if (response.fromCache) {
incrementMetric('cache.hits');
} else {
incrementMetric('cache.misses');
}
},
onError(error) {
incrementMetric('requests.failed');
console.warn('Request failed:', error.config.url, error.status);
},
};
```
> ℹ️ **Note:** This is just an example. You need to implement the `incrementMetric` function yourself to record or report performance metrics as needed in your application.
## 🏷️ Headers
Click to expand
`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
```typescript
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
```typescript
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.
- **`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.
> ⚠️ **Accept-Encoding in Node.js:**
> In Node.js, decompression is handled by the fetch implementation, and users should ensure their environment supports the encodings.
- **`Content-Type`**:
Set automatically based on the request body type:
- For JSON-serializable bodies (objects, arrays, etc.):
`application/json; charset=utf-8`
- For `URLSearchParams`:
`application/x-www-form-urlencoded`
- For `ArrayBuffer`/typed arrays:
`application/octet-stream`
- For `FormData`, `Blob`, `File`, or `ReadableStream`:
**Not set** as the header is handled automatically by the browser and by Node.js 18+ native fetch.
The `Content-Type` header is **never overridden** if you set it manually.
**Summary:**
You only need to set headers manually if you want to override these defaults. Otherwise, `fetchff` will handle the correct headers for most use cases, including advanced scenarios like file uploads, form submissions, and binary data.
## 🌀 Interceptors
Click to expand
Interceptor functions can be provided to customize the behavior of requests and responses. These functions are invoked at different stages of the request lifecycle and allow for flexible handling of requests, responses, and errors.
### Example
```typescript
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);
},
onRetry(response, attempt) {
// Log retry attempts for monitoring and debugging
console.warn(
`Retrying request (attempt ${attempt + 1}):`,
response.config.url,
);
// Modify config for the upcoming retry request
response.config.headers['Authorization'] = 'Bearer your-new-token';
// Log error details for failed attempts
if (response.error) {
console.warn(
`Retry reason: ${response.error.status} - ${response.error.statusText}`,
);
}
// You can implement custom retry logic or monitoring here
// For example, send retry metrics to your analytics service
},
retry: {
retries: 3,
delay: 1000,
backoff: 1.5,
},
});
```
### Configuration
The following options are available for configuring interceptors in the `fetchff` settings:
- **`onRequest(config) => config`**:
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:_ `undefined` (no modification).
- **`onResponse(response) => response`**:
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:_ `undefined` (no modification).
- **`onError(error) => error`**:
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:_ `undefined` (no modification).
- **`onRetry(response, attempt) => response`**:
Type: `RetryInterceptor | RetryInterceptor[]`
A function or an array of functions that are invoked before each retry attempt. Each function receives the response object (containing error information) and the current attempt number as arguments, allowing you to implement custom retry logging, monitoring, or conditional retry logic.
_Default:_ `undefined` (no retry interception).
All interceptors are asynchronous and can modify the provided config or response objects. You don't have to return a value, but if you do, any returned properties will be merged into the original argument.
### Interceptor Execution Order
`fetchff` follows specific execution patterns for interceptor chains:
#### **Request Interceptors: FIFO (First In, First Out)**
Request interceptors execute in the order they are defined - from global to specific:
```typescript
// Execution order: 1 → 2 → 3 → 4
const api = createApiFetcher({
onRequest: (config) => {
/* 1. Global interceptor */
},
endpoints: {
getData: {
onRequest: (config) => {
/* 2. Endpoint interceptor */
},
},
},
});
await api.getData({
onRequest: (config) => {
/* 3. Request interceptor */
},
});
```
#### **Response Interceptors: LIFO (Last In, First Out)**
Response interceptors execute in reverse order - from specific to global:
```typescript
// Execution order: 3 → 2 → 1
const api = createApiFetcher({
onResponse: (response) => {
/* 3. Global interceptor (executes last) */
},
endpoints: {
getData: {
onResponse: (response) => {
/* 2. Endpoint interceptor */
},
},
},
});
await api.getData({
onResponse: (response) => {
/* 1. Request interceptor (executes first) */
},
});
```
This pattern ensures that:
- **Request interceptors** can progressively enhance configuration from general to specific
- **Response interceptors** can process data from specific to general, allowing request-level interceptors to handle the response first before global cleanup or logging
### 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.
## 🌐 Network Revalidation
Click to expand
`fetchff` provides intelligent network revalidation features that automatically keep your data fresh based on user interactions and network connectivity. These features help ensure users always see up-to-date information without manual intervention.
### Focus Revalidation
When `refetchOnFocus` is enabled, requests are automatically triggered when the browser window regains focus (e.g., when users switch back to your tab).
```typescript
const { data } = await fetchf('/api/user-profile', {
refetchOnFocus: true, // Revalidate when window gains focus
cacheTime: 300, // Cache for 5 minutes, but still revalidate on focus
});
```
### Network Reconnection Revalidation
The `refetchOnReconnect` feature automatically revalidates data when the browser detects that internet connectivity has been restored after being offline.
```typescript
const { data } = await fetchf('/api/notifications', {
refetchOnReconnect: true, // Revalidate when network reconnects
cacheTime: 600, // Cache for 10 minutes, but revalidate when back online
});
```
### Adaptive Timeouts
`fetchff` automatically adjusts request timeouts based on connection speed to provide optimal user experience:
```typescript
// Automatically uses:
// - 30 seconds timeout on normal connections
// - 60 seconds timeout on slow connections (2G/3G)
const { data } = await fetchf('/api/data');
// You can still override with custom timeout
const { data: customTimeout } = await fetchf('/api/data', {
timeout: 10000, // Force 10 seconds regardless of connection speed
});
// Check connection speed manually
import { isSlowConnection } from 'fetchff';
if (isSlowConnection()) {
console.log('User is on a slow connection');
// Adjust your app behavior accordingly
}
```
### How It Works
1. **Event Listeners**: `fetchff` automatically attaches global event listeners for `focus` and `online` events when needed
2. **Background Revalidation**: Network revalidation uses background requests that don't show loading states to users
3. **Automatic Cleanup**: Event listeners are properly managed and cleaned up to prevent memory leaks
4. **Smart Caching**: Revalidation works alongside caching - fresh data updates the cache for future requests
5. **Stale-While-Revalidate**: Use `staleTime` to control when background revalidation happens automatically
6. **Connection Awareness**: Automatically detects connection speed and adjusts timeouts for better reliability
### Configuration Options
Both revalidation features can be configured globally or per-request, and work seamlessly with cache timing:
```typescript
import { createApiFetcher } from 'fetchff';
const api = createApiFetcher({
baseURL: 'https://api.example.com',
// You can set all settings globally
refetchOnFocus: true,
refetchOnReconnect: true,
cacheTime: 300, // Cache for 5 minutes
staleTime: 60, // Consider fresh for 1 minute, then background revalidate
endpoints: {
getCriticalData: {
url: '/critical-data',
// Override global settings for specific endpoints
refetchOnFocus: true,
refetchOnReconnect: true,
staleTime: 30, // More aggressive background revalidation for critical data
},
getStaticData: {
url: '/static-data',
// Disable revalidation for static data
refetchOnFocus: false,
refetchOnReconnect: false,
staleTime: 3600, // Background revalidate after 1 hour
},
},
});
```
### Use Cases
**Focus Revalidation** is ideal for:
- Real-time dashboards and analytics
- Social media feeds and chat applications
- Financial data and trading platforms
- Any data that changes frequently while users are away
**Reconnection Revalidation** is perfect for:
- Mobile applications with intermittent connectivity
- Offline-capable applications
- Critical data that must be current when online
- Applications used in areas with unstable internet
### Best Practices
1. **Combine with appropriate cache and stale times**:
```typescript
const { data: notifications } = await fetchf('/api/notifications', {
cacheTime: 600, // Cache for 10 minutes
staleTime: 60, // Background revalidate after 1 minute
refetchOnFocus: true,
});
const { data: userProfile } = await fetchf('/api/profile', {
cacheTime: 1800, // Cache for 30 minutes
staleTime: 600, // Background revalidate after 10 minutes
refetchOnReconnect: true,
});
```
2. **Use `staleTime` for automatic background updates** - Data stays fresh without user interaction:
```typescript
// Good: Automatic background revalidation for dynamic data
const { data: notifications } = await fetchf('/api/notifications', {
cacheTime: 600, // Cache for 10 minutes
staleTime: 60, // Background revalidate after 1 minute
refetchOnFocus: true,
});
// Good: Less frequent updates for semi-static data
const { data: userProfile } = await fetchf('/api/profile', {
cacheTime: 1800, // Cache for 30 minutes
staleTime: 600, // Background revalidate after 10 minutes
refetchOnReconnect: true,
});
```
3. **Use selectively** - Don't enable for all requests to avoid unnecessary network traffic:
```typescript
// Good: Enable for critical, changing data
const { data: userNotifications } = await fetchf('/api/notifications', {
refetchOnFocus: true,
refetchOnReconnect: true,
});
// Avoid: Don't enable for static configuration data
const { data: appConfig } = await fetchf('/api/config', {
cacheTime: 3600, // Cache for 1 hour
staleTime: 0, // Disable background revalidation
refetchOnFocus: false,
refetchOnReconnect: false,
});
```
4. **Consider user experience** - Network revalidation happens silently in the background, providing smooth UX without loading spinners.
> ⚠️ **Browser Support**: These features work in all modern browsers that support the `focus` and `online` events. In server-side environments (Node.js), these options are safely ignored.
>
> **React Native**: Use `setEventProvider()` to enable these features. See the [React Native](#react-native) section for details.
## 🗄️ Cache Management
Click to expand
The caching mechanism in fetchf() and createApiFetcher() enhances performance by reducing redundant network requests and reusing previously fetched data when appropriate. This system ensures that cached responses are managed efficiently and only used when considered "fresh". Below is a breakdown of the key parameters that control caching behavior and their default values.
> ⚠️ **When using in Node.js:**
> Cache and deduplication are in-memory and per-process. For distributed or serverless environments, consider external caching if persistence is needed.
### Example
```typescript
const { data } = await fetchf('https://api.example.com/', {
cacheTime: 300, // Cache is valid for 5 minutes, set -1 for indefinite cache. By default no cache.
cacheKey: (config) => `${config.url}-${config.method}`, // Custom cache key based on URL and method, default automatically generated
cacheBuster: (config) => config.method === 'POST', // Bust cache for POST requests, by default no busting.
skipCache: (response, config) => response.status !== 200, // Skip caching on non-200 responses, by default no skipping
cacheErrors: false, // Cache error responses as well as successful ones, default false
staleTime: 600, // Data is considered fresh for 10 minutes before background revalidation (0 by default, meaning no background revalidation)
});
```
### Configuration
The caching system can be fine-tuned using the following options when configuring the:
- **`cacheTime`**:
Type: `number`
Specifies the duration, in seconds, for which a cache entry is considered "fresh." Once this time has passed, the entry is considered stale and may be refreshed with a new request. Set to -1 for indefinite cache.
_Default:_ `undefined` (no caching).
- **`cacheKey`**:
Type: `CacheKeyFunction | string`
A string or function used to generate a custom cache key for the request cache, deduplication etc. If not provided, a default key is created by hashing various parts of the request, including `Method`, `URL`, query parameters, and headers etc. Providing string can help to greatly improve the performance of the requests, avoid unnecessary request flooding etc.
You can provide either:
- A **string**: Used directly as the cache key for all requests using matching string.
- A **function**: Receives the full request config as an argument and should return a unique string key. This allows you to include any relevant part of the request (such as URL, method, params, body, or custom logic) in the cache key.
**Example:**
```typescript
cacheKey: (config) =>
`${config.method}:${config.url}:${JSON.stringify(config.params)}`;
```
This flexibility ensures you can control cache granularity—whether you want to cache per endpoint, per user, or based on any other criteria.
_Default:_ Auto-generated based on request properties (see below).
- **`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).
- **`cacheErrors`**:
Type: `boolean`
Determines whether error responses (such as HTTP 4xx or 5xx) should also be cached. If set to `true`, both successful and error responses are stored in the cache. If `false`, only successful responses are cached.
_Default:_ `false`.
- **`staleTime`**:
Specifies the time in seconds during which cached data is considered "fresh" before it becomes stale and triggers background revalidation (SWR: stale-while-revalidate).
- Set to a number greater than `0` to enable SWR: cached data will be served instantly, and a background request will update the cache after this period.
- Set to `0` to treat data as stale immediately (always eligible for refetch).
- Set to `undefined` to disable SWR: data is never considered stale and background revalidation is not performed.
_Default:_ `undefined` to disable SWR pattern (data is never considered stale) or `300` (5 minutes) in libraries like React.
### How It Works
1. **Cache Lookup**:
When a request is made, `fetchff` first checks the internal cache for a matching entry using the generated cache key. If a valid and "fresh" cache entry exists (within `cacheTime`), the cached response is returned immediately. If the native `fetch()` option `cache: 'reload'` is set, the internal cache is bypassed and a fresh request is made.
2. **Cache Key Generation**:
Each request is uniquely identified by a cache key, which is auto-generated from the URL, method, params, headers, and other relevant options. You can override this by providing a custom `cacheKey` string or function for fine-grained cache control.
3. **Cache Busting**:
If a `cacheBuster` function is provided, it determines whether to invalidate (bust) the cache for a given request. This is useful for scenarios like forcing fresh data on `POST` requests or after certain actions.
4. **Conditional Caching**:
The `skipCache` function allows you to decide, per response, whether it should be stored in the cache. For example, you can skip caching for error responses (like HTTP 4xx/5xx) or based on custom logic.
5. **Network Request and Cache Update**:
If no valid cache entry is found, or if caching is skipped or busted, the request is sent to the network. The response is then cached according to your configuration, making it available for future requests.
### 🔄 Cache and Deduplication Integration
Understanding how caching works together with request deduplication is crucial for optimal performance:
#### **Cache-First, Then Deduplication**
```typescript
// Multiple components requesting the same data
const userProfile1 = useFetcher('/api/user/123', { cacheTime: 300 });
const userProfile2 = useFetcher('/api/user/123', { cacheTime: 300 });
const userProfile3 = useFetcher('/api/user/123', { cacheTime: 300 });
// Flow:
// 1. First request checks cache → cache miss → network request initiated
// 2. Second request checks cache → cache miss → joins in-flight request (deduplication)
// 3. Third request checks cache → cache miss → joins in-flight request (deduplication)
// 4. When network response arrives → cache is populated → all requests receive same data
```
#### **Cache Hit Scenarios**
```typescript
// First request (cache miss - goes to network)
const request1 = fetchf('/api/data', { cacheTime: 300, dedupeTime: 5000 });
// After 2 seconds - cache hit (no deduplication needed)
setTimeout(() => {
const request2 = fetchf('/api/data', { cacheTime: 300, dedupeTime: 5000 });
// Returns cached data immediately, no network request
}, 2000);
// After 10 minutes - cache expired, new request
setTimeout(() => {
const request3 = fetchf('/api/data', { cacheTime: 300, dedupeTime: 5000 });
// Cache expired → new network request → potential for deduplication again
}, 600000);
```
#### **Deduplication Window vs Cache Time**
- **`dedupeTime`**: Prevents duplicate requests during a short time window (milliseconds)
- **`cacheTime`**: Stores successful responses for longer periods (seconds)
- **Integration**: Deduplication handles concurrent requests, caching handles subsequent requests
```typescript
const config = {
dedupeTime: 2000, // 2 seconds - for rapid concurrent requests
cacheTime: 300, // 5 minutes - for longer-term storage
};
// Timeline example:
// T+0ms: Request A initiated → network call starts
// T+500ms: Request B initiated → joins Request A (deduplication)
// T+1500ms: Request C initiated → joins Request A (deduplication)
// T+2500ms: Request D initiated → deduplication window expired, but cache hit!
// T+6000ms: Request E initiated → cache hit (no network call needed)
```
### ⏰ Understanding staleTime vs cacheTime
The relationship between `staleTime` and `cacheTime` enables sophisticated data freshness strategies:
#### **Cache States and Timing**
```typescript
const fetchWithTimings = fetchf('/api/user-feed', {
cacheTime: 600, // Cache for 10 minutes
staleTime: 60, // Consider fresh for 1 minute before background revalidation
});
// Data lifecycle:
// T+0: Fresh data - served from cache, no background request
// T+30s: Still fresh - served from cache, no background request
// T+90s: Stale but cached - served from cache + background revalidation
// T+300s: Still stale - served from cache + background revalidation
// T+650s: Cache expired - network request required, shows loading state
```
#### **Practical Combinations**
**High-Frequency Updates (Real-time Data)**
```typescript
const realtimeData = {
cacheTime: 30, // Cache for 30 seconds
staleTime: 5, // Fresh for 5 seconds only
// Result: Frequent background updates, always responsive UI
};
```
**Balanced Performance (User Data)**
```typescript
const userData = {
cacheTime: 300, // Cache for 5 minutes
staleTime: 60, // Fresh for 1 minute
// Result: Good performance + reasonable freshness
};
```
**Static Content (Configuration)**
```typescript
const staticConfig = {
cacheTime: 3600, // Cache for 1 hour
staleTime: 1800, // Fresh for 30 minutes
// Result: Minimal network usage for rarely changing data
};
```
#### **Background Revalidation Behavior**
```typescript
// When staleTime expires but cacheTime hasn't:
const { data } = await fetchf('/api/notifications', {
cacheTime: 600, // 10 minutes total cache
staleTime: 120, // 2 minutes of "freshness"
});
// T+0: Returns cached data immediately, no background request
// T+150s: Returns cached data immediately + triggers background request
// T+150s: Background request completes → cache silently updated
// T+650s: Cache expired → full loading state + network request
```
### Auto-Generated Cache Key Properties
By default, `fetchff` generates a cache key automatically using a combination of the following request properties:
| Property | Description | Default Value |
| ----------------- | ----------------------------------------------------------------------------------------- | --------------- |
| `method` | The HTTP method used for the request (e.g., GET, POST). | `'GET'` |
| `url` | The full request URL, including the base URL and endpoint path. | `''` |
| `headers` | Request headers, **filtered to include only cache-relevant headers** (see below). | |
| `body` | The request payload (for POST, PUT, PATCH, etc.), stringified if it's an object or array. | |
| `credentials` | Indicates whether credentials (cookies) are included in the request. | `'same-origin'` |
| `params` | Query parameters serialized into the URL (objects, arrays, etc. are stringified). | |
| `urlPathParams` | Dynamic URL path parameters (e.g., `/user/:id`), stringified and encoded. | |
| `withCredentials` | Whether credentials (cookies) are included in the request. | |
#### Header Filtering for Cache Keys
To ensure stable cache keys and prevent unnecessary cache misses, `fetchff` only includes headers that affect response content in cache key generation. The following headers are included:
**Content Negotiation:**
- `accept` - Affects response format (JSON, HTML, etc.)
- `accept-language` - Affects localization of response
- `accept-encoding` - Affects response compression
**Authentication & Authorization:**
- `authorization` - Affects access to protected resources
- `x-api-key` - Token-based access control
- `cookie` - Session-based authentication
**Request Context:**
- `content-type` - Affects how request body is interpreted
- `origin` - Relevant for CORS or tenant-specific APIs
- `referer` - May influence API behavior
- `user-agent` - Only if server returns client-specific content
**Custom Headers:**
- `x-requested-with` - Distinguishes AJAX requests
- `x-client-id` - Per-client/partner identity
- `x-tenant-id` - Multi-tenant segmentation
- `x-user-id` - Explicit user context
- `x-app-version` - Version-specific behavior
- `x-feature-flag` - Feature rollout controls
- `x-device-id` - Device-specific responses
- `x-platform` - Platform-specific content (iOS, Android, web)
- `x-session-id` - Session-specific responses
- `x-locale` - Locale-specific content
Headers like `user-agent`, `accept-encoding`, `connection`, `cache-control`, tracking IDs, and proxy-related headers are **excluded** from cache key generation as they don't affect the actual response content.
These properties are combined and hashed to create a unique cache key for each request. This ensures that requests with different parameters, bodies, or cache-relevant headers are cached separately while maintaining stable cache keys across requests that only differ in non-essential headers. If that does not suffice, you can always use `cacheKey` (string | function) and supply it to particular requests. You can also build your own `cacheKey` function and simply update defaults to reflect it in all requests. Auto key generation would be entirely skipped in such scenarios.
## 🔁 Deduplication & In-Flight Requests
Click to expand
`fetchff` automatically deduplicates identical requests that are made within a configurable time window, ensuring that only one network request is sent for the same endpoint and parameters. This is especially useful for scenarios where multiple components or users might trigger the same request simultaneously (e.g., rapid user input, concurrent UI updates).
> ⚠️ **When using in Node.js:**
> Request queueing and deduplication are per-process. In multi-process or serverless environments, requests are not deduplicated across instances.
### How Deduplication Works
- When a request is made, `fetchff` checks if an identical request (same URL, method, params, and body) is already in progress or was recently completed within the `dedupeTime` window.
- If such a request exists, the new request will "join" the in-flight request and receive the same response when it completes, rather than triggering a new network call.
- This mechanism reduces unnecessary network traffic and ensures that all consumers receive the same response for identical requests made in quick succession.
### Configuration
- **`dedupeTime`**:
- Type: `number`
- Default: `0` (milliseconds)
- Specifies the time window during which identical requests are deduplicated. If set to `0`, deduplication is disabled.
### Example
```typescript
import { fetchf } from 'fetchff';
// Multiple rapid calls to the same endpoint will be deduplicated
fetchf('/api/search', { params: { q: 'test' }, dedupeTime: 2000 });
fetchf('/api/search', { params: { q: 'test' }, dedupeTime: 2000 });
// Only one network request will be sent within the 2-second window
```
### Benefits
- Prevents duplicate network requests for the same resource.
- Reduces backend load and improves frontend performance.
- Ensures that all consumers receive the same response for identical requests made in quick succession.
This deduplication logic is applied both to standalone `fetchf()` calls and to endpoints created with `createApiFetcher()`.
## ✋ Request Cancellation
Click to expand
fetchff simplifies making API requests by allowing customizable features such as request cancellation, retries, and response flattening. When a new request is made to the same API endpoint, the plugin automatically cancels any previous requests that haven't completed, ensuring that only the most recent request is processed.
It also supports:
- 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
```javascript
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
Click to expand
Polling can be configured to repeatedly make requests at defined intervals until certain conditions are met. This allows for continuously checking the status of a resource or performing background updates.
### Example
```typescript
const { data } = await fetchf('https://api.example.com/', {
pollingInterval: 5000, // Poll every 5 seconds (useful for regular polling at intervals)
pollingDelay: 1000, // Wait 1 second before each polling attempt begins
maxPollingAttempts: 10, // Stop polling after 10 attempts
shouldStopPolling(response, 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. It is useful for regular, periodic polling.
_Default:_ `0` (polling disabled).
- **`pollingDelay`**:
Type: `number`
The time (in milliseconds) to wait before each polling attempt begins. It is useful if you want to throttle or stagger requests, or wait a bit before each poll. It basically adds a delay before each poll is started (including the first one).
_Default:_ `0` (no delay).
- **`maxPollingAttempts`**:
Type: `number`
Maximum number of polling attempts before stopping. Set to `0` or negative number for unlimited attempts.
_Default:_ `0` (unlimited).
- **`shouldStopPolling`**:
Type: `(response: any, attempt: number) => boolean`
A function to determine if polling should stop based on the response, error, or the current polling attempt number (attempt starts with `1`). 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, 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. **Polling Delay**:
The `pollingDelay` setting introduces a delay before each polling attempt, allowing for finer control over the timing of requests.
3. **Maximum Polling Attempts**:
The `maxPollingAttempts` setting limits the number of polling attempts. If the maximum number of attempts is reached, polling stops automatically.
4. **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.
5. **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
Click to expand
The retry mechanism can be used to handle transient errors and improve the reliability of network requests. This mechanism automatically retries requests when certain conditions are met, providing robustness in the face of temporary failures. Below is an overview of how the retry mechanism works and how it can be configured.
### Example
```typescript
const { data } = await fetchf('https://api.example.com/', {
retry: {
retries: 5,
delay: 100, // Override default adaptive delay (normally 1s/2s based on connection)
maxDelay: 5000, // Override default adaptive maxDelay (normally 30s/60s based on connection)
resetTimeout: true, // Resets the timeout for each retry attempt
backoff: 1.5,
retryOn: [500, 503],
// Retry on specific errors or based on custom logic
shouldRetry(response, attempt) {
// Retry if the status text is Not Found (404)
if (response.error && response.error.statusText === 'Not Found') {
return true;
}
// Use `response.data` to access any data from fetch() response
const data = response.data;
// Let's say your backend returns bookId as "none". You can force retry by returning "true".
if (data?.bookId === 'none') {
return true;
}
return attempt < 3; // Retry up to 3 times.
},
},
});
```
In this example, the request will retry only on HTTP status codes 500 and 503, as specified in the `retryOn` array. The `resetTimeout` option ensures that the timeout is restarted for each retry attempt. The custom `shouldRetry` function adds further logic: if the server response contains `{"bookId": "none"}`, a retry is forced. Otherwise, the request will retry only if the current attempt number is less than 3. Although the `retries` option is set to 5, the `shouldRetry` function limits the maximum attempts to 3 (the initial request plus 2 retries).
**Note:** When not overridden, `fetchff` automatically adapts retry delays based on connection speed:
- **Normal connections**: 1s initial delay, 30s max delay
- **Slow connections (2G/3G)**: 2s initial delay, 60s max delay
Additionally, you can handle "Not Found" (404) responses or other specific status codes in your retry logic. For example, you might want to retry when the status text is "Not Found":
```typescript
shouldRetry(response, attempt) {
// Retry if the status text is Not Found (404)
if (response.error && response.error.statusText === 'Not Found') {
return true;
}
// ...other logic
return null; // Fallback to `retryOn` status code check
}
```
This allows you to customize retry behavior for cases where a resource might become available after a short delay, or when you want to handle transient 404 errors gracefully.
The whole Error object is under `response.error` generally.
### 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. **Default is adaptive**: 1 second (1000 ms) for normal connections, 2 seconds (2000 ms) on slow connections (2G/3G). Subsequent retries use an exponentially increasing delay based on the `backoff` parameter.
_Default:_ `1000` / `2000` (adaptive based on connection speed).
- **`maxDelay`**:
Type: `number`
Maximum delay (in milliseconds) between retry attempts. **Default is adaptive**: 30 seconds (30000 ms) for normal connections, 60 seconds (60000 ms) on slow connections (2G/3G). The delay will not exceed this value, even if the exponential backoff would suggest a longer delay.
_Default:_ `30000` / `60000` (adaptive based on connection speed).
- **`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. It means that after the first failure, wait for x seconds. After the second failure, wait for x _ 1.5 seconds. After the third failure, wait for x _ 1.5^2 seconds, and so on.
_Default:_ `1.5`.
- **`resetTimeout`**:
Type: `boolean`
If set to `true`, the timeout for the request is reset for each retry attempt. This ensures that the timeout applies to each individual retry rather than the entire request lifecycle.
_Default:_ `true`.
- **`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
If used in conjunction with `shouldRetry`, the `shouldRetry` function takes priority, and falls back to `retryOn` only if it returns `null`.
- **`shouldRetry(response: FetchResponse, currentAttempt: Number) => boolean`**:
Type: `RetryFunction`
Function that determines whether a retry should be attempted based on the error or successful response (if `shouldRetry` is provided) object, and the current attempt number. This function receives the error object and the attempt number as arguments. The boolean returned indicates decision. If `true` then it should retry, if `false` then abort and don't retry, if `null` then fallback to `retryOn` status codes check.
_Default:_ `undefined`.
### 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`.
- If `resetTimeout` is enabled, the timeout is reset for each retry attempt.
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.
### 429 Retry-After Handling
When a request receives a **429 Too Many Requests** response, `fetchff` will automatically check for the `Retry-After` header and use its value to determine the delay before the next retry attempt. This works for both seconds and HTTP-date formats, and falls back to your configured delay if the header is missing or invalid.
**How it works:**
- If the server responds with 429 and a `Retry-After` header, the delay for the next retry will be set to the value from that header (in ms).
- If the header is missing or invalid, the default retry delay is used.
**Example:**
```typescript
const { data } = await fetchf('https://api.example.com/', {
retry: {
retries: 2,
delay: 1000, // fallback if Retry-After is missing
retryOn: [429], // 429 is already checked by default so it is not necessary to add it
},
});
```
If the server responds with:
```
HTTP/1.1 429 Too Many Requests
Retry-After: 5
```
The next retry will wait 5000ms before attempting again.
If the header is an HTTP-date, the delay will be calculated as the difference between the date and the current time.
## 🧩 Response Data Transformation
Click to expand
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. It returns functions like `response.json()` that can be called idempotently without throwing errors or consuming the underlying stream multiple times. You can also use `parser` option to fully control returned response (useful for response streaming).
### **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`, images, video, audio, pdf, zip):** Parses the response as an `ArrayBuffer`. Use `response.blob()` to get a `Blob`, or `response.bytes()` to get a `Uint8Array`.
- **URL-encoded Form Data (`application/x-www-form-urlencoded`):** Parses the response as `FormData`.
- **Text (`text/*`):** Parses the response as plain text.
- **Other/Custom types (XML, CSV, etc.):** Returns raw text. Use the `parser` option for custom parsing.
If the `Content-Type` header is missing or not recognized, the plugin defaults to returning text. If the text looks like JSON (starts with `{` or `[`), it will be auto-parsed as JSON.
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.
> ⚠️ **When using in Node.js:**
> In Node.js, using FormData, Blob, or ReadableStream may require additional polyfills or will not work unless your fetch polyfill supports them.
### Custom `parser` Option
You can provide a custom `parser` function to override the default content-type based parsing. This is useful for formats like XML, CSV, or any proprietary format. The `parser` can be set globally (via `createApiFetcher()` or `setDefaultConfig()`) or per-request.
```typescript
import { fetchf } from 'fetchff';
// Example: Parse XML responses
const { data } = await fetchf('/api/data.xml', {
async parser(response) {
const text = await response.text();
return new DOMParser().parseFromString(text, 'application/xml');
},
});
// Example: Parse CSV responses
const { data: csvData } = await fetchf('/api/report.csv', {
async parser(response) {
const text = await response.text();
return text.split('\n').map((row) => row.split(','));
},
});
```
You can also set it globally:
```typescript
import { createApiFetcher } from 'fetchff';
const api = createApiFetcher({
apiUrl: 'https://example.com/api',
parser: async (response) => {
const text = await response.text();
return new DOMParser().parseFromString(text, 'application/xml');
},
endpoints: {
getReport: { url: '/report' },
},
});
```
### `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
Click to expand
Every request returns a standardized response object from native fetch() extended by a few handful properties:
### Response Object Structure
```typescript
interface FetchResponse<
ResponseData = any,
RequestBody = any,
QueryParams = any,
PathParams = any,
> extends Response {
data: ResponseData | null; // The parsed response data, or null/defaultResponse if unavailable
error: ResponseError<
ResponseData,
RequestBody,
QueryParams,
PathParams
> | null; // Error details if the request failed, otherwise null
config: RequestConfig; // The configuration used for the request
status: number; // HTTP status code
statusText: string; // HTTP status text
headers: HeadersObject; // Response headers as a key-value object
isSuccess: boolean; // True if request is successful (2xx status codes).
isError: boolean; // True if the response contains an error
}
```
- **`data`**:
The actual data returned from the API, or `null`/`defaultResponse` if not available.
- **`error`**:
An object containing error details if the request failed, or `null` otherwise. Includes properties such as `name`, `message`, `status`, `statusText`, `request`, `config`, and the full `response`.
- **`config`**:
The complete configuration object used for the request, including URL, method, headers, and parameters.
- **`status`**:
The HTTP status code of the response (e.g., 200, 404, 500).
- **`statusText`**:
The HTTP status text (e.g., 'OK', 'Not Found', 'Internal Server Error').
- **`headers`**:
The response headers as a plain key-value object.
- **`isSuccess`**:
Indicates whether the request was successful (2xx status codes).
- **`isError`**:
True if the response is an error (non-2xx status code, network error, or request failed).
The whole response of the native `fetch()` is attached as well.
Error object in `error` looks as follows:
- **Type**: `ResponseError | null`
- An object with details about any error that occurred or `null` otherwise.
- **`name`**: The name of the error, that is `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.
- **`isCancelled`**: A boolean property on the error object indicating whether the request was cancelled before completion
## 🔍 Error Handling
Click to expand
Error handling strategies define how to manage errors that occur during requests. You can configure the strategy option to specify what should happen when an error occurs. This affects whether promises are rejected, if errors are handled silently, or if default responses are provided. You can also combine it with onError interceptor for more tailored approach.
The native `fetch()` API function doesn't throw exceptions for HTTP errors like `404` or `500` — it only rejects the promise if there is a network-level error (e.g. the request fails due to a DNS error, no internet connection, or CORS issues). The `fetchf()` function brings consistency and lets you align the behavior depending on chosen strategy. By default, all errors are rejected.
### Configuration
#### `strategy`
**`reject`**: (default)
Promises are rejected, and global error handling is triggered. You must use `try/catch` blocks to handle errors.
```typescript
import { fetchf } from 'fetchff';
try {
const { data } = await fetchf('https://api.example.com/users', {
strategy: 'reject', // Default strategy - can be omitted
timeout: 5000,
});
console.log('Users fetched successfully:', data);
} catch (error) {
// Handle specific error types
if (error.status === 404) {
console.error('API endpoint not found');
} else if (error.status >= 500) {
console.error('Server error:', error.statusText);
} else {
console.error('Request failed:', error.message);
}
}
```
**`softFail`**:
Returns a response object with additional property of `error` when an error occurs and does not throw any error. This approach helps you to handle error information directly within the response's `error` object without the need for `try/catch` blocks.
> ⚠️ **Always Check the error Property:**
> When using the softFail or defaultResponse strategies, the promise will not throw on error.
> You must always check the error property in the response object to detect and handle errors.
```typescript
import { fetchf } from 'fetchff';
const { data, error } = await fetchf('https://api.example.com/users', {
strategy: 'softFail',
timeout: 5000,
});
if (error) {
// Handle errors without try/catch
console.error('Request failed:', {
status: error.status,
message: error.message,
url: error.config?.url,
});
// Show user-friendly error message
if (error.status === 429) {
console.log('Rate limited. Please try again later.');
} else if (error.status >= 500) {
console.log('Server temporarily unavailable. Please try again.');
}
} else {
console.log('Users fetched successfully:', data);
}
```
Check `Response Object` section below to see how `error` object is structured.
**`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.
> ⚠️ **Always Check the error Property:**
> When using the softFail or defaultResponse strategies, the promise will not throw on error.
> You must always check the error property in the response object to detect and handle errors.
```typescript
import { fetchf } from 'fetchff';
const { data, error } = await fetchf(
'https://api.example.com/user-preferences',
{
strategy: 'defaultResponse',
defaultResponse: {
theme: 'light',
language: 'en',
notifications: true,
},
timeout: 5000,
},
);
if (error) {
console.warn('Failed to load user preferences, using defaults:', data);
// Log error for debugging but continue with default values
console.error('Preferences API error:', error.message);
} else {
console.log('User preferences loaded:', data);
}
// Safe to use data regardless of error state
document.body.className = data.theme;
```
**`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.
> ⚠️ **When using in Node.js:**
> The 'silent' strategy will hang the promise forever. Use with caution, especially in backend/server environments.
```typescript
async function myLoadingProcess() {
const { data } = await fetchf('https://api.example.com/', {
strategy: 'silent',
});
// In case of an error nothing below will ever be executed.
console.log('Th