fquery 1.5.0-beta.1

fquery is a powerful async state management solution for Flutter. It caches, updates, and fully manages asynchronous data in your Flutter apps.

⚡Are you ready to supercharge your Flutter app development?

Introducing fquery - an easy-to-use, yet efficient and reliable asynchronous state management solution for Flutter! It effortlessly caches, updates, and fully manages asynchronous data in your Flutter apps.

With this powerful tool at your disposal, managing server state (REST API, GraphQL, etc), local databases like SQLite, or anything async has never been easier. Just provide a Future and watch the magic unfold.

Community

Trusted & Used by

The project's growth has almost been completely organic, it has grown popular in the developer community and is growing by the day, consider starring it if you've found it useful.

As a developer, you too can leverage the power of this tool to create a high-quality mobile application that provides an exceptional user experience. So, why not choose it for your next project and take advantage of its powerful features to deliver a seamless experience to your users?

🌌 Features

• Easy to use
• Powerful and fully customizable
• No boilerplate code required
• Data fetching logic agnostic
• Automatic caching and garbage collection
• Automatic re-fetching of stale data
• State data invalidation
• Manual updates available
• Dependent queries
• Parallel queries
• Infinite queries
• Mutations

❔Defining the problem

Have you ever wondered how to effectively manage server state in your Flutter apps? Many developers resort to using Riverpod, Bloc, `FutureBuilder``, or any other general-purpose state management solution. However, these solutions often lead to writing repetitive code that handles data fetching, caching, and other logic.

The truth is, that general-purpose state management solutions are not the best choice when it comes to handling asynchronous server state. This is due to the unique nature of server state - it is asynchronous and requires specific APIs for fetching and updating. Additionally, the server state is stored in a remote location, which means it can be modified without your knowledge from anywhere in the world. This alone requires a lot of effort to keep the data synchronized and ensure that it is up-to-date.

How does ⚡fquery tackle this problem?

fquery is powered by flutter_hooks. It is very similar to swr and react-query. With fquery, you can make use of easy-to-use hooks to retrieve data from a Future and the rest of the process is automated. fquery is highly configurable, allowing you to customize it to meet your specific needs. You can configure every aspect of it to make it work optimally for your use case.

📄 Example

Here's a very simple widget that makes use of the useQuery hook:

class Posts extends HookWidget {
  const Posts({Key? key}) : super(key: key);

  @override
  Widget build(BuildContext context) {
    final posts = useQuery(['posts'], getPosts);

    return Builder(
      builder: (context) {
        if (posts.isLoading) {
          return const Center(child: CircularProgressIndicator());
        }

        if (posts.isError) {
          return Center(child: Text(posts.error!.toString()));
        }

        return ListView.builder(
          itemCount: posts.data!.length,
          itemBuilder: (context, index) {
            final post = posts.data![index];
            return ListTile(
              title: Text(post.title),
            );
          },
        );
      },
    );
  }
}

🧑‍💻 Usage

You can either install flutter_hooks before using this library or use the widgets like QueryBuilder and MutationBuilder that comes with fquery. You'll need to wrap your entire app inside a QueryClientProvider and you are good to go.

void main() {
  runApp(
    QueryClientProvider(
      queryClient: queryClient,
      child: CupertinoApp(

Queries

To query data in your widgets, you can either extend your widget using HookWidget or StatefulHookWidget (for stateful widgets) or use the QueryBuilder widget, see below. These classes are exported from the flutter_hooks package.

A query instance is a subscription to asynchronous data stored in the cache. Every query needs -

• A Query key, uniquely identifies the query stored in the cache.
• A Future that either resolves or throws an error

The same query key can be used in multiple instances of the useQuery hook and the data will be shared throughout the app.

Future<List<Post>> getPosts() async {
  final res = await Dio().get('https://jsonplaceholder.typicode.com/posts');
  return (res.data as List)
      .map((e) => Post.fromJson(e as Map<String, dynamic>))
      .toList();
}

class Posts extends HookWidget {
  const Posts({Key? key}) : super(key: key);

  @override
  Widget build(BuildContext context) {
    final posts = useQuery(['posts'], getPosts);

The returned value of the useQuery hook is an instance of UseQueryResult and contains all the information related to that query. A Builder widget comes in handy when rendering the results.

// The query has no data to display
if (posts.isLoading) {
  return const Center(child: CircularProgressIndicator());
}

// An error has occurred
if (posts.isError) {
  return Center(child: Text(posts.error!.toString()));
}

// Success, data is ready to display
return ListView.builder(
  itemCount: posts.data!.length,
  itemBuilder: (context, index) {
    final post = posts.data![index];
    return ListTile(
      title: Text(post.title),
    );
  },
);

Query without flutter_hooks

You can have queries without extending your widget with HookWidget. Just use the QueryBuilder widget and you're good to go.

The QueryBuilder takes 3 required arguments, first one is the query key, second one is query function and the third one is a named parameter, builder.

QueryBuilder<List<Todo>, Error>(
  const ['todos'],
  todosAPI.getAll,
  refetchOnMount: RefetchOnMount.never,
  refetchInterval: const Duration(seconds: 10),
  enabled: isEnabled.value,
  builder: (context, todos) {
    if (todos.isLoading) {
      return const Center(
        child: CupertinoActivityIndicator(),
      );
    }
    if (todos.isError) {
      return Center(
        child: Text(todos.error.toString()),
      );
    }

    ...

Query configuration

A query is fully customizable to match your needs, these configurations can be passed as named parameters into the useQuery hook

// These are default configurations
final posts = useQuery(
  ['posts'],
  getPosts,
  enabled: true,
  cacheDuration: const Duration(minutes: 5),
  refetchInterval: null // The query will not re-fetch by default,
  refetchOnMount: RefetchOnMount.stale,
  staleDuration: const Duration(seconds: 10),
  retryCount: 3,
  retryDelay: const Duration(seconds: 1, milliseconds: 500)
);

• enabled - specifies if the query fetcher function is automatically called when the widget renders and can be used for dependent queries.
• cacheDuration - specifies the duration unused/inactive cache data remains in memory; the cached data will be garbage collected after this duration. The longest duration will be used when different values are specified in multiple instances of the query.
• refetchInterval - specifies the time interval in which all queries will re-fetch the data, setting it to null (default) will turn off re-fetching.
• refetchOnMount - specifies the behavior of the query instance when the widget is first built and the data is already available.
  • RefetchOnMount.always - will always re-fetch when the widget is built.
  • RefetchOnMount.stale - will fetch the data if it is stale (see staleDuration).
  • RefetchOnMount.never - will never re-fetch.
• staleDuration - specifies the duration until the data becomes stale. This value applies to each query instance individually.
• retryCount - specifies the number of times the query will retry before showing an error
• retryDelay - specifies the delay between each retry

Dependent Query

A dependent query is a query that depends on another variable for execution, or even any other query. Probably you want to run a query only after some other query, or data in a query that you don't have, e.g. a Future, or to fetch data only when a variable takes a certain value, e.g. a bool like isAuthenticated, for all of this or similar, dependent query can ease your load. To use this, simply pass the enabled option.

final user = useQuery(['users', email], getUserByEmail);

// This query will not execute until the above is successful and the username is available
final username = user.data?.username;
final posts = useQuery(['posts', ], getPosts, enabled: !username);


final isAuthenticated = session != null;
final keys = useQuery(['keys', session.id], enabled: isAuthenticated)

Infinite queries

Infinite scroll is a very common UI pattern and fquery comes with a useInfiniteQuery hook to handle those needs. In addition to queryKey and queryFn, it requires an initialPageParam and getNextPageParam option. The query function receives the pageParam parameter that can be used to fetch the current page.

It can also be used to create bi-directional infinite scroll by using the getPreviousPageParam.

Example:

final items = useInfiniteQuery<PageResult, Error, int>(
  ['infinity'],
  (page) => infinityAPI.get(page),
  initialPageParam: 1,
  getNextPageParam: ((lastPage, allPages, lastPageParam, allPageParam) {
    return lastPage.hasMore ? lastPage.page + 1 : null;
  }),
);

Parallel queries

Parallel queries are queries that are executed in parallel. When the number of parallel queries does not change, there is no extra effort to use parallel queries.

// These will execute in parallel
final posts = useQuery(['posts'], getProfile)
final comments = useQuery(['comments'], getProfile)

Dynamic Parallel queries

If the number of queries you need to execute is changing from render to render, you cannot use manual querying since that would violate the rules of hooks. Instead fquery provides the useQueries hook for that purpose.

// See how the number of queries are changing with `text.value`
final posts = useQueries<Post, Error>(
  List<UseQueriesOptions<Post, Error>>.generate(text.value,
  (i) => UseQueriesOptions(
    queryKey: ['posts', i + 1],
    fetcher: () => getPost(i + 1),
    refetchOnMount: RefetchOnMount.never,
  ),
));

Global fetching indicators

If you want to know the number of all the queries that are being fetched at the moment, you can use the useIsFetching hook provided by the library.

// `fetchingCount` is an int
final fetchingCount = useIsFetching();

Query invalidation

This technique can manually mark the cached data as stale and potentially even re-fetch them. This is especially useful when you know that the data has been changed. QueryClient (see below) has an invalidateQueries() method that allows you to do that. You can make use of the useQueryClient hook to obtain the instance of QueryClient that you passed with QueryClientProvider.

final queryClient = useQueryClient();

// Invalidate every query with a key that starts with `post`
queryClient.invalidateQueries(['posts']);

// here, both queries will be invalidated
final posts = useQuery(['posts'], getPosts);
final post = useQuery(['posts', 1], getPosts);


// Use `exact: true` to exactly match the query
queryClient.invalidateQueries(['posts'], exact: true);

// here, only this will invalidate
final posts = useQuery(['posts'], getPosts);

When a query is invalidated, two things will happen:

• It marks it as stale and this overrides any staleDuration configuration passed to useQuery.
• If the query is being used in a widget, it will be re-fetched, otherwise, it will be re-fetched when it is used by a widget at a later point in time.

Manual updates

You probably already know how the data is changed and don't want to re-fetch the whole data again. You can set it manually using the setQueryData() method on the QueryClient. It takes a query key and an updater function. If the query data doesn't exist already in the cache (that's why previous is nullable), it'll be created.

final queryClient = useQueryClient();

// The `Type` of returned data must match the `Type` of data
// stored in the cache, otherwise an error will be thrown
queryClient.setQueryData<List<Post>>(['posts'], (previous) {
  return previous?.map((post) {
    return post.copyWith(
      title: "lorem ipsum"
    );
  }).toList() ?? <Post>[]
})

QueryClient

A QueryClient is used to interact with the query cache. It is made available throughout the app using a QueryClientProvider. It can be configured to change the default configurations of the queries.

final queryClient = QueryClient(
  defaultQueryOptions: DefaultQueryOptions(
    cacheDuration: Duration(minutes: 20),
    refetchInterval: Duration(minutes: 5),
    refetchOnMount: RefetchOnMount.always,
    staleDuration: Duration(minutes: 3),
  ),
);

void main() {
  runApp(
    QueryClientProvider(
      queryClient: queryClient,
      child: CupertinoApp(

QueryClient.removeQueries

This method is used to remove any query from the cache. If the query's data is currently being rendered on the screen then it will still show and the query will also be removed from the cache.

QueryClientBuilder widget

There's also a builder widget where you can get access to the query client.

QueryClientBuilder(
  builder: (context, queryClient) {
    return MutationBuilder(
      (id) async {
        await todosAPI.delete(todo.id);
        return id;
      },
      onSuccess: (id, _, ctx) {
        queryClient.setQueryData<List<Todo>>(
          ['todos'],
          (previous) {
            if (previous == null) return [];
            return previous.where((e) {

Mutations

Similar to queries, you can also use the useMutation hook to mutate data on the server or just anywhere, just return a Future in your mutation function and you're good to go.

The following example illustrates almost the full usage of the features that come with mutations. Here we're adding a new todo asynchronously and also doing optimistic updates.

Example

final addTodoMutation = useMutation<Todo, Exception, String, List<Todo>>(
        todosAPI.add, onMutate: (text) async {
      final previousTodos =
          queryClient.getQueryData<List<Todo>>(['todos']) ?? [];

      // Optimistically update the todo list
      queryClient.setQueryData<List<Todo>>(['todos'], (previous) {
        final id = Random().nextInt(pow(10, 6).toInt());
        final newTodo = Todo(id: id, text: text);
        return [...(previous ?? []), newTodo];
      });

      // Pass the original data as context to the next functions
      return previousTodos;
    }, onError: (err, text, previousTodos) {
      // On failure, revert to original data
      queryClient.setQueryData<List<Todo>>(
        ['todos'],
        (_) => previousTodos as List<Todo>,
      );
    }, onSettled: (data, error, variables, ctx) {
      // Refetch the query anyway (either error or success)
      // Or we can manually add the returned todo (result) in the onSuccess callback
      client.invalidateQueries(['todos']);
      todoInputController.clear();
    });

Usage

To use mutations, you need a mutation function that will receive a variable parameter when you call the mutate function. Here in the example, it's the text parameter that we're using as a variable that the mutation function will receive.

Like queries, to use mutations you can either extend the widget using HookWidget or StatefulHookWidget (for stateful widgets) or use the MutationBuilder widget, see below.

The useMutation hook takes 4 type arguments -

• TData - type of data that'll be returned from the mutation function.
• TError - type of error that'll be thrown when the mutation fails.
• TVariables - type of the variable that your mutation function will receive.
• TContext - type of the context object you'll pass around in mutation callbacks. It has been illustrated in the example how onMutate returns the original list of todos to revert when the mutation fails.

You can also pass callback functions like onSuccess or onError -

• onMutate - this callback will be called before the mutation is executed and is passed with the same variables the mutation function would receive.
• onSuccess - this callback will be called if the mutation was successful and receives the result of the mutation as an argument (in addition to the passed variables in the mutation function).
• onError - this callback will be called if the mutation wasn't successful and receives the error as an argument (in addition to the passed variables in the mutation function).
• onSettled - this callback will be called after the mutation has been executed and will receive both the result (if successful) and error(if unsuccessful), in case of success the error will be null and vice-versa.

The useMutation hook will return [UseMutationResult] which has everything associated with the mutation.

final TData? data;
final TError? error;
final bool isIdle;
final bool isPending;
final bool isSuccess;
final bool isError;
final MutationStatus status;
final Future<void> Function(TVariables) mutate;
final DateTime? submittedAt;
final void Function() reset;
final TVariables? variables;

Mutation without flutter_hooks

You can have queries without extending your widget with HookWidget. Just use the QueryBuilder widget and you're good to go.

The MutationBuilder takes 2 required arguments, first one is the mutation function and second one is a named parameter, builder.

MutationBuilder((id) async {
  await todosAPI.delete(todo.id);
  return id;
}, onSuccess: (id, _, ctx) {
  client.setQueryData<List<Todo>>(
    ['todos'],
    (previous) {
      if (previous == null) return [];
      return previous.where((e) {
        return (e.id != id);
      }).toList();
    },
  );
}, builder: (context, mutation) {
  return CupertinoButton(
  ...

Contributing

If you've ever wanted to contribute to open source, and a great cause, now is your chance ✨, feel free to open an issue or submit a PR at the GitHub repo. See Contribution guide for more details.

Contributors ✨

Thanks go to these wonderful people:

Piyush 🐛 💻 📖 🎨 🚧 👀

Cynthia 📖

Rajvir Singh 🎨

Kollin Murphy 📖 💻

Tuco T. 🐛 💻 🤔

tsuyoshi wada 🐛 💻

du-nt 🤔

This project follows the all-contributors specification. Contributions of any kind are welcome!