npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

spa-sm

v1.6.0

Published

Single Page Application State Manager - 'A bit too much structure' for single page web apps

Downloads

12

Readme

Spa-sm - Single Page Application State Management

Spasm provides "just enough structure" to navigate and manage state in single page web applications.

Concepts

Spasm's goal is to help you organize data flow, and separate concerns that should not be conflated:

  • Data fetching
  • Data processing
  • Managing client-side state
  • Rendering
  • Triggering actions
  • Syncing/responding to URL updates

Spasm defines three main concepts to help with this separation.

Routes

Spasm includes a simple router. It supports URL templates with named parts, and allows you to order routes by precedence. A route is associated with a page name, and is used to resolve which page is being requested.

The router is bi-directional. It can resolve page name and parameters from URLs, and it can generate URLs from a page name and parameters. The router also parses and exposes query string parameters, but cannot route based on them.

Pages

A page is an object with at least a render method, and optionally a method to fetch data, and a method to process data. When a page is requested, Spasm will perform the following steps to render the page:

  • Get data with page.getData({state, location})
  • Optionally seed state for that page through page.seedState({pageData, state, location})
  • Prepare data for rendering with page.prepareData({pageData, state, location})
  • Finalize data for rendering with app.finalizeData(preparedData, location, state)
  • Call on the app's render function with page.render and the prepared data

Spasm makes no assumptions about your render function, what or who will render your app, or how this happens.

The goal of this pipeline is to avoid putting data fetching and processing inside UI components. By keeping these separate in pure functions, the UI becomes a 1:1 visual representation of a data structure. This data structure is dramatically easier to test than a complex and ever-changing UI.

The details of the various function calls can be found in the API documentation below.

Actions

When the user interacts with the application, it needs to perform some work in response, and most of the time re-render. Spasm suggests that an action is an event with some data. This means that the only thing event handlers in e.g. React components need to do is to emit an event, which again means less logic inside the components. Spasm even goes as far as suggesting that the component shouldn't even know what arguments to pass along with the action. This decision is made by prepareData (more on this below).

An action is implemented as a function. In addition to performing application-specific work, actions can load a new page, manipulate the query parameters of the current page, manipulate the client-side state, or trigger a refresh.

An example

We start by creating an app instance. Doing so will not make anything appear on screen. The app is bound to an element on the page, and takes in a list of routes:

import {createApp} from 'spa-sm';
import react from 'react';

const app = createApp({
  render(component, data) {
    react.render(component(data), document.getElementById('app'));
  }
});

The app needs at least one page (we'll add an edit page later):

const {h1, div, p} = React.DOM;

const UserComponent = React.createFactory(React.createClass({
  render() {
    return h1({}, 'Hello world');
  }
}));

app.addPage('viewUser', '/users/:id', {
  render: UserComponent
});

Finally, a call to start will trigger the initial render:

app.start();

Visiting http://localhost:10666/users/chris should now display "Hello world!". Let's display the requested user's name. To do so, we will add a function to process the data so the component only receives the bits it needs to render.

const UserComponent = React.createFactory(React.createClass({
  render() {
    return h1({}, this.props.name);
  }
}));

app.addPage('viewUser', '/users/:id', {
  prepareData({location: {params: {id}}}) {
    const name = id[0].toUpperCase() + id.slice(1);
    return {
      name,
      title: `User: ${name}`
    };
  },

  render: UserComponent
});

Including a title in the data returned from prepareData will cause Spasm to update document.title to reflect it.

In practice, most pages need to fetch some data from somewhere. Data fetch can be synchronous or asynchronous. If it is asynchronous, getData should return a promise. The getData function receives the current client-side state and the location, which it can use to inform its retrieval. For instance, if the app requires logging in a user, you might want to store the user in the client-side state, and look it up to retrieve user-specific data:

const UserComponent = React.createFactory(React.createClass({
  render() {
    return div({},
               h1({}, this.props.name),
               p({}, this.props.info));
  }
}));

app.addPage('viewUser', '/users/:id', {
  getData({location: {params: {id}}, state}) {
    return {
      id,
      name: id[0].toUpperCase() + id.slice(1),
      info: 'Some old user'
    };
  },

  prepareData({pageData: user, location}) {
    return {
      name: user.name,
      title: `User: ${user.name}!`,
      info: user.info
    };
  },

  render: UserComponent
});

The data returned from getData (or eventually resolved from a promise returned from it) will be included in the object passed to prepareData as pageData. prepareData pulls relevant data from the different sources and produces a single object that will be rendered by the component.

Initial state can be provided when creating the app:

const app = createApp(document.getElementById('app'), {
  state: {currentUser: 'Christian'},

  render(component, data) {
    react.render(component(data), document.getElementById('app'));
  }
});

Let's add an action to the application - a regular link that loads a different page. First, we will define the action itself:

app.addAction('gotoURL', url => app.gotoURL(url));

To trigger this action in the component, we should first expose it in prepareData:

prepareData({pageData: user, location}) {
  const editUserURL = app.getURL('editUser', user);
  return {
    name: user.name,
    title: `User: ${user.name}!`,
    info: user.info,
    editUserURL,
    actions: {
      edit: ['gotoURL', editUserURL]
    }
  };
}

This adds two things: the editUserURL, which is generated from the bi-directional router, and the action. Note how prepareData makes all the decisions on what data to display and what data to pass with actions. This separation of concerns leaves our components wonderfully oblivious about details in our data structures, and tightly focused on rendering:

const UserComponent = React.createFactory(React.createClass({
  render() {
    return div({},
               h1({}, this.props.name),
               p({}, this.props.info),
               p({}, a({
                 href: this.props.editUserURL,
                 onClick(e) {
                   e.preventDefault()
                   app.triggerAction(this.props.actions.edit);
                 }
               }, 'Edit user')));
  }
}));

The event handler used with the click event is so common, that the app instance provides a helper for it:

a({
  href: this.props.editPageURL,
  onClick: app.performAction(this.props.actions.edit)
}, 'Edit page')

This will currently fail, because we haven't implemented the editUser page yet, here's a stub:

const EditUserComponent = React.createFactory(React.createClass({
  render() {
    return div({},
               h1({}, 'Edit user'));
  }
}));

app.addPage('editUser', '/users/:id/edit', {
  render: EditUserComponent
});

For our final example, we will consider adding a "flash message" - a message pinned to the top of the page regardless of which URL you're currently watching. We will start by adding a new action. It will add some data to the client-side state and trigger a re-render.

app.addAction('updateState', (state, data) => app.updateState(state));

Next up, we will include an action for the UI so it can trigger the flash:

prepareData({pageData: user, location}) {
  const editUserURL = app.getURL('editUser', user);
  return {
    name: user.name,
    title: `User: ${user.name}!`,
    info: user.info,
    editUserURL,
    actions: {
      edit: ['gotoURL', editUserURL],
      triggerFlash: ['updateState', {flash: {message: 'I am a flash'}}]
    }
  };
}

Because the flash needs to be available on all pages, we don't want to repeat the code in prepareData on all the pages. This is what the app's global finalizeData is for. It receives the data as prepared by the rendering page, along with the client-side state and the current location:

import assign from 'lodash/object/assign';
import pick from 'lodash/object/pick';

const app = createApp(document.getElementById('app'), {
  state: {currentUser: 'Christian'},

  finalizeData(data, location, state) {
    return assign(data, pick(state, 'flash'));
  },

  render(component, data) {
    react.render(component(data), document.getElementById('app'));
  }
});

Now every page in the app will be able to access flash on the data passed to render (if there is a flash). Finally, we render it:

const FlashMessage = React.createFactory(React.createClass({
  render() {
    if (!this.props.message) {
      return null;
    }
    return p({className: 'flash-message'}, this.props.message);
  }
}));

const UserComponent = React.createFactory(React.createClass({
  render() {
    console.log(this.props);
    const {flash, name, info, editUserURL, actions} = this.props;

    return div({},
               FlashMessage(flash),
               h1({}, name),
               p({}, info),
               p({}, a({
                 href: editUserURL,
                 onClick: app.performAction(actions.edit)
               }, 'Edit user')),
               p({}, button({
                 onClick: app.performAction(actions.triggerFlash)
               }, 'Trigger flash')));
  }
}));

const EditUserComponent = React.createFactory(React.createClass({
  render() {
    return div({},
               FlashMessage(this.props.flash),
               h1({}, 'Edit user'));
  }
}));

You might want the flash to disappear automatically after some time. To do so, replace updateState(state) with flashState(state, ms), which will undo the state changes after ms milliseconds.

That about sums up the important bits about Spasm. Below you will find complete API docs for all the details.

The demo can be found in the demo directory, and can be run with npm start (following npm install).

API docs

const app = createApp({state, finalizeData, render})

Creates a new application instance.

state

Initial client-side state, as an object. Optional.

finalizeData(preparedData, location, state)

A function that finalizes data processing before every render. preparedData is the data returned from the current page's prepareData function. location is the result of the router's parsing, see below. state is the client-side state.

The object returned from this function will be passed to the page's render function.

render(component, data)

A function that renders the app given the current state. It is passed the current page's render function (which, when using React, should be a React component factory), and the prepared data.

app.loadURL(url[, state])

Use the router to resolve the page for this URL, fetch its data, prepare it, and render. Optionally update client-side state as well.

app.gotoURL(url[, state])

Like loadURL, but also push the url to the browser.

app.replaceURL(url[, state])

Like gotoURL, but will replace the current URL.

app.triggerAction(action[, arg1, arg2])

Triggers an action. If trying to trigger an action that has no handlers, this will throw an exception. The action is an array, where the first element is the name of the action, and the remaining elements will be passed as arguments to the action. You can also pass additional arguments at call time.

The action will be called with the action arguments, call-time arguments (if any), and the current data as the last argument, e.g.:

app.addAction('test', (actionArg1, actionArg2, callTimeArg1, {location, pageData, state}) => {
  // ...
});

// ...

app.triggerAction(['test', {action: 'args'}, 42], 13);

app.getURL(pageName, params)

Uses the bi-directional router to generate a URL from the page name and parameters.

app.getCurrentURL()

Returns the URL corresponding to the current page and current state. You can modify query in location and have it reflected in the URL returned from getCurrentURL.

app.action('updateQueryParams', (params, {location}) => {
  app.updateQueryParams(params);
});

app.refresh([state])

Runs the current page over again - fetches data, prepares data and renders. Optionally takes some new client-side state.

app.updateState(state)

Add some local state and re-render the application.

app.updateQueryParams(params)

Add some URL query parameters. This will cause the app to refresh with the new query parameters included in the URL. The new URL will also be loaded in the browser.

app.addAction(actionName, handler)

Define an action. There's no expectations on how the action should behave. Most actions will want to cause visual changes at some point, and for this purpose, you might be interested in the app's loadURL, updateState, updateQueryParams, and refresh functions.

The action handler will be called with an action-specific argument as its first argument, and the current data (which consists of {location, state, pageData}) as its second argument.

app.performAction(action)

Convenience for use as event handler in React components. Returns a function that can be used as an event handler. The handler will call preventDefault and then trigger the action.

React.DOM.a({onClick: app.performAction(['gotoURL', '/'])}, 'Home');

app.addPage(pageName, route, {getData, prepareData, render}, {prefix})

Add a page definition. The prefix can be used for static URL prefixes that you don't want to be part of the URLs.

pageName

A string name for the page. Used with routing.

route

The URL template that matches this page.

app.start()

Kick things off by rendering location.href.

Pages

Pages are plain old JavaScript objects. Every page must implement the render function, which can be a single React component, e.g.: {render: SomeComponent}. Additionally, pages may implement getData and prepareData.

page.getData({state, location})

Fetch data for this page. May return data directly, or a promise that resolves with the data. The state contains whatever was previously put in it. location is described with the router below.

page.prepareData({pageData, state, location})

Perform any data processing necessary for rendering the UI and return the object to render. pageData is the data retrieved with getData.

If the object returned from prepareData includes a title, it will be used for the page title (e.g. document.title).

page.seedState({pageData, state, location})

Seed state for this page. This function is called after getData and before prepareData, and allows you to add state without triggering recursive renders (as updateState would do, if called from getData or prepareData). The return value from this function is merged into the state before prepareData is called. prepareData will receive the freshly updated state.

The router

The router is not exposed directly, although it is completely possible to use it on its own if desired.

const routingTable = createRoutes(routes[, {prefix}])

Creates a routing table. routes is an array of routes, where each route is a tuple of [pageName, urlTemplate]. The prefix can be used to ignore static URL prefixes, e.g. to target a URL like /my/app/news/23 with the route /news/:id and the prefix /my/app.

const url = toURLString({path, query})

Produces a URL string from a path and a query parameter object.

const url = getURL(routes, page, params)

Generates the URL to the given page type with the given parameters.

const location = getLocation(routes, url)

Returns the matching page from the routing table for the giving URL. If no route matches, it returns null. The location description includes the following properties:

String page

The name of the matching page.

String url

The URL matched against.

String path

Only the path part of the URL matched against.

Object params

Parameters matched from the route. Any :paramName placeholder from URL templates are included. For instance, if the matching route was /stuff/:id, then params will contain the id property.

Object query

The query string, parsed into an object.