Functional Reactive Programming and HTML

WebSharper.UI is a library providing a novel, pragmatic and convenient approach to UI reactivity. It includes:

  • An HTML library usable both from the server side and from the client side, which you can use to build HTML pages either by calling F# functions to create elements, or by instantiating template HTML files.
  • A reactive layer for expressing user inputs and values computed from them as time-varying values. This approach is related to Functional Reactive Programming (FRP). This reactive system integrates with the HTML library to create reactive documents. If you are familiar with Facebook React, then you will find some similarities with this approach: instead of explicitly inserting, modifying and removing DOM nodes, you return a value that represents a DOM tree based on inputs. The main difference is that these inputs are nodes of the reactive system, rather than a single state value associated with the component.
  • Client-side routing using the same endpoint type declaration as WebSharper server-side routing.

This page is an overview of the capabilities of WebSharper.UI. You can also check the full reference of all the API types and modules.

Get the package from NuGet: WebSharper.UI.

Using HTML

WebSharper.UI's core type for HTML construction is Doc. A Doc can represent a single DOM node (element, text), but it can also be a sequence of zero or more nodes. This allows you to treat equally any HTML snippet that you want to insert into a document, whether it consists of a single element or not.

Additionally, client-side Docs can be reactive. A same Doc can consist of different elements at different moments in time, depending on user input or other variables. See the reactive section to learn more about this.

Constructing HTML

Docs

The main means of creating Docs is by using the functions in the WebSharper.UI.Html module. Every HTML element has a dedicated function, such as div or p, which takes a sequence of attributes (of type Attr) and a sequence of child nodes (of type Doc). Additionally, the text function creates a text node.

open WebSharper.UI.Html

let myDoc =
    div [] [
        h1 [] [ text "Functional Reactive Programming and HTML" ]
        p [] [ text "WebSharper.UI is a library providing a novel, pragmatic and convenient approach to UI reactivity. It includes:" ]
        ul [] [
            li [] [ text "..." ]
        ]
    ]

Result:

<div>
  <h1>Functional Reactive Programming and HTML</h1>
  <p>WebSharper.UI is a library providing a novel, pragmatic and convenient
     approach to UI reactivity. It includes:</p>
  <ul>
    <li>...</li>
  </ul>
</div>

Some HTML tags, such as option, collide with standard library names and are therefore only located in the Tags submodule.

let myDropdown =
    select [] [
        Tags.option [] [ text "First choice" ]
        Tags.option [] [ text "Second choice" ]
        Tags.option [] [ text "Third choice" ]
    ]

Result:

<select>
  <option>First choice</option>
  <option>Second choice</option>
  <option>Third choice</option>
</select>

One thing to note is that the tag functions described above actually return a value of type Elt, which is a subtype of Doc that is guaranteed to always consist of exactly one element and provides additional APIs such as Dom to get the underlying Dom.Element. This subtyping means that you will sometimes need to upcast the result of such a function with :> Doc to appease the compiler; you can see an example of this below in the example for Doc.Empty.

Additional functions in the Doc can create or combine Docs:

  • Doc.Empty creates a Doc consisting of zero nodes. This can be useful for example when you may not need to insert an element depending on a condition.

    let myForm (withDropdown: bool) =
        form [] [
            input [ attr.name "name" ] []
            (if withDropdown then myDropdown :> Doc else Doc.Empty)
        ]
    

    Result:

    <form>
      <input name="name" />
    </form>
    
    or:
    
    <form>
      <input name="name" />
      <!-- ... contents of myDropdown here ... -->
    </form>
    
  • Doc.Append creates a Doc consisting of the concatenation of two Docs.

    let titleAndBody =
        Doc.Append
            (h1 [] [ text "Functional Reactive Programming and HTML" ])
            (p [] [ text "WebSharper.UI is a library providing ..." ])
    

    Result:

    <h1>Functional Reactive Programming and HTML</h1>
    <p>WebSharper.UI is a library providing ...</p>
    

For the mathematically enclined, the functions Doc.Empty and Doc.Append make Docs a monoid.

  • Doc.Concat generalizes Append by concatenating a sequence of Docs.

    let thisPage =
        Doc.Concat [
            h1 [] [ text "Functional Reactive Programming and HTML" ]
            p [] [ text "WebSharper.UI is a library providing ..." ]
            ul [] [
                li [] [ text "..." ]
            ]
        ]
    

    Result:

    <h1>Functional Reactive Programming and HTML</h1>
    <p>WebSharper.UI is a library providing ...</p>
    <ul>
      <li>...</li>
    </ul>
    
  • Doc.Element creates an element with the given name, attributes and children. It is equivalent to the function with the same name from the Html module. This function is useful if the tag name is only known at runtime, or if you want to create a non-standard element that isn't available in Html. The following example creates a header tag of a given level (h1, h2, etc).

    let makeHeader (level: int) (content: string) =
        Doc.Element ("h" + string level) [] [ text content ]
    

    Result:

    <h1>content...</h1>
    or
    <h2>content...</h2>
    or etc.
    
  • Doc.Verbatim creates a Doc from plain HTML text.
    Security warning: this function does not perform any checks on the contents, and can be a code injection vulnerability if used improperly. We recommend avoiding it unless absolutely necessary, and properly sanitizing user inputs if you do use it. If you simply want to use HTML syntax instead of F# functions, take a look at templating.

    let plainDoc =
        Doc.Verbatim "
            <h1 onclick=\"alert('And it is unsafe!')\">
                This is plain HTML!
            </h1>"
    

    Result:

    <h1 onclick="alert('And it is unsafe!')">
        This is plain HTML!
    </h1>
    

Attrs

To create attributes, use corresponding functions from the attr submodule.

let myFormControl =
    select [ attr.name "mySelect" ] [
        Tags.option [ attr.value "first" ] [ text "First choice" ]
        Tags.option [ attr.value "second" ] [ text "Second choice" ]
        Tags.option [
            attr.value "third"
            attr.selected "selected"
        ] [ text "Third choice" ]
    ]

Result:

<select name="mySelect">
  <option value="first">First choice</option>
  <option value="second">Second choice</option>
  <option value="third" selected="selected">Third choice</option>
</select>

Some attributes, notably class and type, are also F# keywords, so they need to be wrapped in double backquotes.

let myMain =
    div [ attr.``class`` "main" ] [ text "..." ]

Result:

<div class="main">...</div>

HTML5 also defines any attribute whose names starts with data- as a valid custom attribute. You can create such an attribute using the function data- from module attr (backquoted since it contains a non-standard character).

let myEltWithData =
    div [ attr.``data-`` "uid" "myDiv" ] [ text "..." ]

Result:

<div data-uid="myDiv">...</div>

Like Doc, a value of type Attr can represent zero, one or more attributes. The functions in the Attr module can create such non-singleton attributes.

  • Attr.Empty creates an empty attribute. This can be useful for example when you may not need to insert an attribute depending on a condition.

    let makeInput (initialValue: option<string>) =
        let valueAttr =
            match initialValue with
            | Some v -> attr.value v
            | None -> Attr.Empty
        input [ valueAttr ] []
    

    Result:

    <input value="initialValue..." />
    or
    <input />
    
  • Attr.Append combines two attributes.

    let passwordAttr =
        Attr.Append (attr.``type`` "password") (attr.placeholder "Password")
    

    Result:

    type="password" placeholder="Password"
    
  • Attr.Concat combines a sequence of attributes.

    let passwordAttr =
        Attr.Concat [
            attr.``type`` "password"
            attr.placeholder "Password"
            attr.``class`` "pw-input"
        ]
    

    Result:

    type="password" placeholder="Password" class="pw-input"
    
  • Attr.Create creates a single attribute. It is equivalent to the function with the same name from the attr module. This function is useful if the attribute name is only known at runtime, or if you want to create a non-standard attribute that isn't available in attr.

    let eltWithNonStandardAttr =
        div [ Attr.Create "my-attr" "my-value" ] [ text "..." ]
    

    Result:

    <div my-attr="my-value">...</div>
    

Event handlers

A special kind of attribute is event handlers. They can be created using functions from the on submodule.

let myButton =
    button [ on.click (fun el ev -> JS.Alert "Hi!") ] [ text "Click me!" ]

The handler function takes two arguments:

  • The element itself, as a native Dom.Element;
  • The triggered event, as a native Dom.Event.
let myButton =
    button [
        attr.id "my-button"
        on.click (fun el ev ->
            JS.Alert (sprintf "You clicked %s at x = %i, y = %i." 
                        el.Id ev.ClientX ev.ClientY)
        )
    ] [ text "Click me!" ]

In addition to the standard HTML events, on.afterRender is a special handler that is called by WebSharper after inserting the element into the DOM.

HTML on the client

To insert a Doc into the document on the client side, use the Doc.Run* family of functions from the module WebSharper.UI.Client. Each of these functions has two variants: one directly taking a DOM Element or Node, and the other suffixed with ById taking the id of an element as a string.

  • Doc.Run and Doc.RunById insert a given Doc as the child(ren) of a given DOM element. Note that it replaces the existing children, if any.

    open WebSharper.JavaScript
    open WebSharper.UI
    open WebSharper.UI.Client
    open WebSharper.UI.Html
    
    let Main () =
        div [] [ text "This goes into #main." ]
        |> Doc.RunById "main"
    
        p [] [ text "This goes into the first paragraph with class my-content." ]
        |> Doc.Run (JS.Document.QuerySelector "p.my-content")
    
  • Doc.RunAppend and Doc.RunAppendById insert a given Doc as the last child(ren) of a given DOM element.

  • Doc.RunPrepend and Doc.RunPrependById insert a given Doc as the first child(ren) of a given DOM element.

  • Doc.RunAfter and Doc.RunAfterById insert a given Doc as the next sibling(s) of a given DOM node.

  • Doc.RunBefore and Doc.RunBeforeById insert a given Doc as the previous sibling(s) of a given DOM node.

  • Doc.RunReplace and Doc.RunReplaceById insert a given Doc replacing a given DOM node.

HTML on the server

On the server side, using sitelets, you can create HTML pages from Docs by passing them to the Body or Head arguments of Content.Page.

open WebSharper.Sitelets
open WebSharper.UI
open WebSharper.UI.Html

let MyPage (ctx: Context<EndPoint>) =
    Content.Page(
        Title = "Welcome!",
        Body = [
            h1 [] [ text "Welcome!" ]
            p [] [ text "This is my home page." ]
        ]
    )

By opening WebSharper.UI.Server, you can also just pass a full page to Content.Page. This is particularly useful together with templates.

let MyPage (ctx: Context<EndPoint>) =
    Content.Page(
        html [] [
            head [] [ title [] [ text "Welcome!" ] ]
            body [] [
                h1 [] [ text "Welcome!" ]
                p [] [ text "This is my home page." ]
            ]
        ]
    )

To include client-side elements inside a page, use the client method, from inside WebSharper.UI.Html.

[<JavaScript>]
module Client =

    let MyControl() =
        button [ on.click (fun el ev -> JS.Alert "Hi!") ] [ text "Click me!" ]

module Server =

    let MyPage (ctx: Context<EndPoint>) =
        Content.Page(
            Title = "Welcome!",
            Body = [
                h1 [] [ text "Welcome!" ]
                p [] [ client <@ Client.MyControl() @> ]
            ]
        )

HTML Templates

WebSharper.UI's syntax for creating HTML is compact and convenient, but sometimes you do need to include a plain HTML file in a project. It is much more convenient for designing to have a .html file that you can touch up and reload your application without having to recompile it. This is what Templates provide. Templates are HTML files that can be loaded by WebSharper.UI, and augmented with special elements and attributes that provide additional functionality:

  • Declaring Holes for nodes, attributes and event handlers that can be filled at runtime by F# code;
  • Declaring two-way binding between F# Vars and HTML input elements (see reactive);
  • Declaring inner Templates, smaller HTML widgets within the page, that can be instantiated dynamically.

All of these are parsed from HTML at compile time and provided as F# types and methods, ensuring that your templates are correct.

Basics

To declare a template, use the Template type provider from the namespace WebSharper.UI.Templating.

open WebSharper.UI.Templating

type MyTemplate = Template<"my-template.html">

To instantiate it, call your type's constructor and then its .Doc() method.

my-template.html:

<div>
  <h1>Welcome!</h1>
  <p>Welcome to my site.</p>
</div>
open WebSharper.UI.Templating

type MyTemplate = Template<"my-template.html">

let myPage = MyTemplate().Doc()

equivalent to:

let myPage =
    div [] [
        h1 [] [ text "Welcome!" ]
        p [] [ text "Welcome to my site." ]
    ]

Note that the template doesn't have to be a full HTML document, but can simply be a snippet or sequence of snippets. This is particularly useful to build a library of widgets using inner templates.

If the template comprises a single HTML element, then an additional method .Elt() is available. It is identical to .Doc(), except its return value has type Elt instead of Doc.

You can also declare a template from multiple files at once using a comma-separated list of file names. In this case, the template for each file is a nested class named after the file, truncated of its file extension.

my-template.html:

<div>
  <h1>Welcome!</h1>
  <p>Welcome to my site.</p>
</div>

second-template.html:

<div>
  <h2>This is a section.</h2>
  <p>And this is its content.</p>
</div>
open WebSharper.UI.Templating

type MyTemplate = Template<"my-template.html, second-template.html">

let myPage =
    Doc.Concat [
        MyTemplate.``my-template``().Doc()
        MyTemplate.``second-template``().Doc()
    ]

equivalent to:

let myPage =
    Doc.Concat [
        div [] [
            h1 [] [ text "Welcome!" ]
            p [] [ text "Welcome to my site." ]
        ]
        div [] [
            h2 [] [ text "This is a section." ]
            p [] [ text "And this is its content." ]
        ]
   ]

Holes

You can add holes to your template that will be filled by F# code. Each hole has a name. To fill a hole in F#, call the method with this name on the template instance before finishing with .Doc().

  • ${HoleName} creates a string hole. You can use it in text or in the value of an attribute.

    my-template.html:

    <div style="background-color: ${Color}">
      <h1>Welcome, ${Name}!</h1>
      <!-- You can use the same hole name multiple times,
           and they will all be filled with the same F# value. -->
      <p>This div's color is ${Color}.</p>
    </div>
    
    let myPage =
        MyTemplate()
            .Color("red")
            .Name("my friend")
            .Doc()
    

    Result:

    <div style="background-color: red">
      <h1>Welcome, my friend!</h1>
      <!-- You can use the same hole name multiple times,
           and they will all be filled with the same F# value. -->
      <p>This div's color is red.</p>
    </div>
    

    On the client side, this hole can also be filled with a View<string> (see reactive) to include dynamically updated text content.

  • The attribute ws-replace creates a Doc or seq<Doc> hole. The element on which this attribute is set will be replaced with the provided Doc(s). The name of the hole is the value of the ws-replace attribute.

    my-template.html:

    <div>
      <h1>Welcome!</h1>
      <div ws-replace="Content"></div>
    </div>
    
    let myPage =
        MyTemplate()
            .Content(p [] [ text "Welcome to my site." ])
            .Doc()
    

    Result:

    <div>
      <h1>Welcome!</h1>
      <p>Welcome to my site.</p>
    </div>
    
  • The attribute ws-hole creates a Doc or seq<Doc> hole. The element on which this attribute is set will have its contents replaced with the provided Doc(s). The name of the hole is the value of the ws-hole attribute.

    my-template.html:

    <div>
      <h1>Welcome!</h1>
      <div ws-hole="Content"></div>
    </div>
    
    let myPage =
        MyTemplate()
            .Content(p [] [ text "Welcome to my site." ])
            .Doc()
    

    Result:

    <div>
      <h1>Welcome!</h1>
      <div>
          <p>Welcome to my site.</p>
      </div>
    </div>
    
  • The attribute ws-attr creates an Attr or seq<Attr> hole. The name of the hole is the value of the ws-attr attribute.

    my-template.html:

    <div ws-attr="MainDivAttr">
      <h1>Welcome!</h1>
      <p>Welcome to my site.</p>
    </div>
    
    let myPage =
        MyTemplate()
            .MainDivAttr(attr.``class`` "main")
            .Doc()
    

    Result:

    <div class="main">
      <h1>Welcome!</h1>
      <p>Welcome to my site.</p>
    </div>
    
  • The attribute ws-var creates a Var hole (see reactive) that is bound to the element. It can be used on the following elements:

    • <input>, <textarea>, <select>, for which it creates a Var<string> hole.
    • <input type="number">, for which it creates a hole that can be one of the following types: Var<int>, Var<float>, Var<CheckedInput<int>>, Var<CheckedInput<float>>.
    • <input type="checkbox">, for which it creates a Var<bool> hole.

    The name of the hole is the value of the ws-var attribute. Text ${Hole}s with the same name can be used, and they will dynamically match the value of the Var.

    my-template.html:

    <div>
      <input ws-var="Name" />
      <div>Hi, ${Name}!</div>
    </div>
    
    let myPage =
        let varName = Var.Create ""
        MyTemplate()
            .Name(varName)
            .Doc()
    

    Result:

    <div class="main">
      <input />
      <div>Hi, [value of above input]!</div>
    </div>
    

    If you don't fill the hole (ie you don't call .Name(varName) above), the Var will be implicitly created, so ${Name} will still be dynamically updated from the user's input.

  • The attribute ws-onclick (or any other event name instead of click) creates an event handler hole of type TemplateEvent -> unit. The argument of type TemplateEvent has the following fields:

    • Target: Dom.Element is the element itself.
    • Event: Dom.Event is the event triggered.
    • Vars has a field for each of the Vars associated to ws-vars in the template.

    my-template.html:

    <div>
      <input ws-var="Name" />
      <button ws-onclick="Click">Ok</button>
    </div>
    
    let myPage =
        MyTemplate()
            .Click(fun t -> JS.Alert("Hi, " + t.Vars.Name.Value))
            .Doc()
    

Filling holes

There are two ways to fill the content for a given hole.

  • The recommended way is by using the method with the hole's name on the template instance, as used in the examples above.

    let myPage =
        MyTemplate()
            .Color("red")
            .Name("my friend")
            .Doc()
    
  • If you need to decide which hole to fill at runtime, you can use the method .With(holeName, content). It will throw a runtime error if the content's type doesn't match the hole's type.

    let myPage =
        MyTemplate()
            .With("Color", "red")
            .With("Name", "my friend")
            .Doc()
    
  • You can of course mix and match both styles.

    let myPage =
        MyTemplate()
            .Color("red")
            .With("Name", "my friend")
            .Doc()
    

Inner templates

To create a template for a widget (as opposed to a full page), you can put it in its own dedicated template file, but another option is to make it an inner template. An inner template is a smaller template declared inside a template file using the following syntax:

  • The ws-template attribute declares that its element is a template whose name is the value of this attribute.
  • The ws-children-template attribute declares that the children of its element is a template whose name is the value of this attribute.

Inner templates are available in F# as a nested class under the main provided type.

my-template.html:

<div ws-attr="MainAttr">
  <div ws-replace="InputFields"></div>
  <div ws-template="Field" class="field-wrapper">
    <label for="${Id}">${Which} Name: </label>
    <input ws-var="Var" placeholder="${Which} Name" name="${Id}" />
  </div>
</div>
type MyTemplate = Template<"my-template.html">

let inputField (id: string) (which: string) (var: Var<string>) =
    MyTemplate.Field()
        .Id(id)
        .Which(which)
        .Var(var)
        .Doc()

let myForm =
    let firstName = Var.Create ""
    let lastName = Var.Create ""
    MyTemplate()
        .MainAttr(attr.``class`` "my-form")
        .InputFields(
            [
                inputField "first" "First" firstName
                inputField "last" "Last" lastName
            ]
        )
        .Doc()

Result:

<div class="my-form">
  <div class="field-wrapper">
    <label for="first">First Name: </label>
    <input placeholder="First Name" name="first" />
  </div>
  <div class="field-wrapper">
    <label for="last">Last Name: </label>
    <input placeholder="Last Name" name="last" />
  </div>
</div>

Instantiating templates in HTML

You can also instantiate a template within another template, entirely in HTML, without the need for F# to glue them together.

A node named <ws-TemplateName> instantiates the inner template TemplateName from the same file. A node named <ws-fileName.TemplateName> instantiates the inner template TemplateName from the file fileName. The file name is the same as the generated class name, so with file extension excluded.

Child elements of the <ws-*> fill holes. These elements are named after the hole they fill.

  • ${Text} holes are filled with the text content of the element.
  • ws-hole and ws-replace holes are filled with the HTML content of the element.
  • ws-attr holes are filled with the attributes of the element.
  • Other types of holes cannot be directly filled like this.

Additionally, attributes on the <ws-*> element itself define hole mappings. That is to say, <ws-MyTpl Inner="Outer"> fills the hole named Inner of the template MyTpl with the value of the hole Outer of the containing template. As a shorthand, <ws-MyTpl Attr> is equivalent to <ws-MyTpl Attr="Attr">.

Any holes that are neither mapped by an attribute nor filled by a child element are left empty.

The following example is equivalent to the example from Inner Templates:

my-template.html:

<div ws-attr="MainAttr">
  <!-- Instantiate the template for input fields. -->
  <!-- Creates the holes FirstVar and SecondVar for the main template. -->
  <!-- Fills the holes Id, Which and Var of Field in both instantiations. -->
  <ws-Field Var="FirstVar">
    <Id>first</Id>
    <Which>First</Which>
  </ws-field>
  <ws-Field Var="SecondVar">
    <Id>last</Id>
    <Which>Last</Which>
  </ws-field>
</div>
<!-- Declare the template for input fields -->
<div ws-template="Field" class="field-wrapper">
  <label for="${Id}">${Which} Name: </label>
  <input ws-var="Var" placeholder="${Which} Name" name="${Id}" />
</div>
type MyTemplate = Template<"my-template.html">

let myForm =
    let firstName = Var.Create ""
    let lastName = Var.Create ""
    MyTemplate()
        .FirstVar(firstName)
        .SecondVar(lastName)
        .Doc()

Controlling the loading of templates

The type provider can be parameterized to control how its contents are loaded both on the server and the client. For example:

type MyTemplate = 
    Template<"my-template.html", 
        clientLoad = ClientLoad.Inline,
        serverLoad = ServerLoad.WhenChanged>

The possible values for clientLoad are:

  • ClientLoad.Inline (default): The template is included in the compiled JavaScript code, and any change to my-template.html requires a recompilation to be reflected in the application.

  • ClientLoad.FromDocument: The template is loaded from the DOM. This means that my-template.html must be the document in which the code is run: either directly served as a Single-Page Application, or passed to Content.Page in a Client-Server Application.

    In this mode, it doesn't make sense for client-side code to instantiate the full template, since you are already inside the document. But the following are possible:

The possible values for serverLoad are:

  • ServerLoad.WhenChanged (default): The runtime sets up a file watcher on the template file, and reloads it whenever it is edited.

  • ServerLoad.Once: The template file is loaded on first use and never reloaded.

  • ServerLoad.PerRequest: The template file is reloaded every time it is needed. We recommend against this option for performance reasons.

Binding directly to the DOM

When using a template from the client side that is declared with clientLoad = ClientLoad.FromDocument, you can directly bind content, Vars, etc. to the DOM. Instead of calling .Doc() to create a Doc, use .Bind(), which returns unit, to just apply the template to the current document.

index.html:

<html>
  <head>
    <title>Welcome!</title>
  </head>
  <body>
    <h1>Welcome!</h1>
    <div ws-replace="Paragraph"></div>
    <button ws-onclick="ClickMe">${ClickText}</button>
  </body>
</html>
type Index = Template<"index.html", ClientLoad.FromDocument>

Index()
    .Paragraph(p [] [text "Welcome to my site."])
    .ClickMe(fun _ -> JS.Alert "Clicked!")
    .ClickText("Click me!")
    .Bind()

Result:

<html>
  <head>
    <title>Welcome!</title>
  </head>
  <body>
    <h1>Welcome!</h1>
    <p>Welcome to my site.</p>
    <button>Click me!</button>
  </body>
</html>

Note that for Bind() to work correctly, the holes need to be present in the document itself. This is not a problem if your project is an SPA. But you can also serve the page from a Sitelet, using the same template on the server side. You can fill some holes on the server side and leave some to be filled by the client side. However, by default, the server-side engine removes unfilled holes from the served document. This is correct behavior in most cases, but here, the client does need the unfilled hole markers like ws-replace or ${...} to be present. So this behavior can be overridden by the optional boolean argument keepUnfilled of the .Doc() and .Elt() methods.

index.html:

<html>
  <head>
    <title>Welcome!</title>
  </head>
  <body>
    <h1>Welcome!</h1>
    <div ws-replace="Paragraph"></div>
    <button ws-onclick="ClickMe">${ClickText}</button>
  </body>
</html>
type Index = Template<"index.html", ClientLoad.FromDocument>

[<JavaScript>]
module Client =

    let Startup() =
        Index()
            .ClickMe(fun _ -> JS.Alert "Clicked!")
            .ClickText("Click me!")
            .Bind()

module Server =
    open WebSharper.UI.Server

    let MyPage() =
        Content.Page(
            Index()
                .Paragraph(p [] [text "Welcome to my site."])
                .Elt(keepUnfilled = true)
                .OnAfterRender(fun _ -> Client.Startup())
        )

Served page:

<html>
  <head>
    <title>Welcome!</title>
  </head>
  <body>
    <h1>Welcome!</h1>
    <p>Welcome to my site.</p>
    <button ws-onclick="ClickMe">${ClickText}</button>
  </body>
</html>

Result after Client.Startup() has run:

<html>
  <head>
    <title>Welcome!</title>
  </head>
  <body>
    <h1>Welcome!</h1>
    <p>Welcome to my site.</p>
    <button>Click me!</button>
  </body>
</html>

Accessing the template's model

Templates allow you to access their "model", ie the set of all the reactive Vars that are bound to it, whether passed explicitly or automatically created for its ws-vars. It is accessible in two ways:

  • In event handlers, it is available as the Vars property of the handler argument.

  • From outside the template: instead of finishing the instanciation of a template with .Doc(), you can call .Create(). This will return a TemplateInstance with two properties: Doc, which returns the template itself, and Vars, which contains the Vars. This method is only available when instantiating the template from the client side.

    my-template.html:

    <div>
      <input ws-var="Name" />
      <div>Hi, ${Name}!</div>
    </div>
    
    let myInstance = MyTemplate().Create()
    myInstance.Vars.Name <- "John Doe"
    let myDoc = myInstance.Doc
    

    Result:

    <div>
      <input value="John Doe" />
      <div>Hi, John Doe!</div>
    </div>
    

Mixing client code in server-side templates

It is possible to include some client-side functionality when creating a template on the server side.

  • If you use ws-var="VarName", the corresponding Var will be created on the client on page startup. However, passing a Var using .VarName(myVar) is not possible, since it would be a server-side Var.

  • Event handlers (such as ws-onclick="EventName") work fully if you pass an anonymous function: .EventName(fun e -> ...). The body of this function will be compiled to JavaScript. You can also pass a top-level function, in this case it must be declared with [<JavaScript>].

    This also includes ws-onafterrender, which causes the given function to be called on page startup.

Special holes in server-side templates

In a server-side template, you must specify the location of where WebSharper can include its generated content. Three special placeholders are provided to include client-side content in the page:

  • scripts is replaced with the JavaScript files required by the client-side code included in the page (including WebSharper-generated .js files). Usage: <script ws-replace="scripts"></script>
  • styles is replaced with the CSS files required by the client-side code included in the page. Usage: <link ws-replace="styles" />
  • meta is replaced with a <meta> tag that contains initialization data for client-side controls. Usage: <meta ws-replace="meta" />

The scripts hole is necessary for correct working of the served page if it contains any client-side WebSharper functionality. The other two are optional: if neither styles nor meta is provided explicilty, then they are included automatically above the content for scripts.

Dynamic templates

It is also possible to create a template without the compile-time safety of the type provider. This is done using the type DynamicTemplate.

This type can be used similarly to Template<...>, with the following limitations:

  • It is server-side only.
  • Its constructor must receive the HTML source as a string.
  • Holes can only be filled with .With(holeName, content).
  • The final instantiation must be done with .Doc().
let myPage =
    DynamicTemplate("""<div style="background-color: ${Color}">Welcome, ${Name}!</div>""")
        .With("Color", "red")
        .With("Name", "my friend")
        .Doc()

Reactive layer

WebSharper.UI's reactive layer helps represent user inputs and other time-varying values, and define how they depend on one another.

Vars

Reactive values that are directly set by code or by user interaction are represented by values of type Var<'T>. Vars are similar to F# ref<'T> in that they store a value of type 'T that you can get or set using the Value property. But they can additionally be reactively observed or two-way bound to HTML input elements.

The following are available from WebSharper.UI.Client:

  • Doc.Input creates an <input> element with given attributes that is bound to a Var<string>.

    let varText = Var.Create "initial value"
    let myInput = Doc.Input [ attr.name "my-input" ] varText
    

    With the above code, once myInput has been inserted in the document, getting varText.Value will at any point reflect what the user has entered, and setting it will edit the input.

  • Doc.IntInput and Doc.FloatInput create an <input type="number"> bound to a Var<CheckedInput<_>> of the corresponding type (int or float). CheckedInput provides access to the validity and actual user input, it is defined as follows:

    type CheckedInput<'T> =
        | Valid of value: 'T * inputText: string
        | Invalid of inputText: string
        | Blank of inputText: string
    
  • Doc.IntInputUnchecked and Doc.FloatInputUnchecked create an <input type="number"> bound to a Var<_> of the corresponding type (int or float). They do not check for the validity of the user's input, which can cause wonky interactions. We recommend using Doc.IntInput or Doc.FloatInput instead.

  • Doc.InputArea creates a <textarea> element bound to a Var<string>.

  • Doc.PasswordBox creates an <input type="password"> element bound to a Var<string>.

  • Doc.CheckBox creates an <input type="checkbox"> element bound to a Var<bool>.

  • Doc.CheckBoxGroup also creates an <input type="checkbox">, but instead of associating it with a simple Var<bool>, it associates it with a specific 'T in a Var<list<'T>>. If the box is checked, then the element is added to the list, otherwise it is removed.

    type Color = Red | Green | Blue
    
    // Initially, Green and Blue are checked.
    let varColor = Var.Create [ Blue; Green ]
    
    let mySelector =
        div [] [
            label [] [
                Doc.CheckBoxGroup [] Red varColor
                text " Select Red"
            ]
            label [] [
                Doc.CheckBoxGroup [] Green varColor
                text " Select Green"
            ]
            label [] [
                Doc.CheckBoxGroup [] Blue varColor
                text " Select Blue"
            ]
        ]
    

    Result:

    <div>
      <label><input type="checkbox" /> Select Red</label>
      <label><input type="checkbox" checked /> Select Green</label>
      <label><input type="checkbox" checked /> Select Blue</label>
    </div>
    

    Plus varColor is bound to contain the list of ticked checkboxes.

  • Doc.Select creates a dropdown <select> given a list of values to select from. The label of every <option> is determined by the given print function for the associated value.

    type Color = Red | Green | Blue
    
    // Initially, Green is checked.
    let varColor = Var.Create Green
    
    // Choose the text of the dropdown's options.
    let showColor (c: Color) =
        sprintf "%A" c
    
    let mySelector =
        Doc.Select [] showColor [ Red; Green; Blue ] varColor
    

    Result:

    <select>
      <option>Red</option>
      <option>Green</option>
      <option>Blue</option>
    </select>
    

    Plus varColor is bound to contain the selected color.

  • Doc.Radio creates an <input type="radio"> given a value, which sets the given Var to that value when it is selected.

    type Color = Red | Green | Blue
    
    // Initially, Green is selected.
    let varColor = Var.Create Green
    
    let mySelector =
        div [] [
            label [] [
                Doc.Radio [] Red varColor
                text " Select Red"
            ]
            label [] [
                Doc.Radio [] Green varColor
                text " Select Green"
            ]
            label [] [
                Doc.Radio [] Blue varColor
                text " Select Blue"
            ]
        ]
    

    Result:

    <div>
      <label><input type="radio" /> Select Red</label>
      <label><input type="radio" checked /> Select Green</label>
      <label><input type="radio" /> Select Blue</label>
    </div>
    

    Plus varColor is bound to contain the selected color.

More variants are available in the Doc module.

Views

The full power of WebSharper.UI's reactive layer comes with Views. A View<'T> is a time-varying value computed from Vars and from other Views. At any point in time the view has a certain value of type 'T.

One thing important to note is that the value of a View is not computed unless it is needed. For example, if you use View.Map, the function passed to it will only be called if the result is needed. It will only be run while the resulting View is included in the document using one of these methods. This means that you generally don't have to worry about expensive computations being performed unnecessarily. However it also means that you should avoid relying on side-effects performed in functions like View.Map.

In pseudo-code below, [[x]] notation is used to denote the value of the View x at every point in time, so that [[x]] = [[y]] means that the two views x and y are observationally equivalent.

Note that several of the functions below can be used more concisely using the V shorthand.

Creating and combining Views

The first and main way to get a View is using the View property of Var<'T>. This retrieves a View that tracks the current value of the Var.

You can create Views using the following functions and combinators from the View module:

  • View.Const creates a View whose value is always the same.

    let v = View.Const 42
    
    // [[v]] = 42
    
  • View.ConstAnyc is similar to Const, but is initialized asynchronously. Until the async returns, the resulting View is uninitialized.

  • View.Map takes an existing View and maps its value through a function.

    let v1 : View<string> = // ...
    let v2 = View.Map (fun s -> String.length s) v1
    
    // [[v2]] = String.length [[v1]]
    
  • View.Map2 takes two existing Views and map their value through a function.

    let v1 : View<int> = // ...
    let v2 : View<int> = // ...
    let v3 = View.Map2 (fun x y -> x + y) v1 v2
    
    // [[v3]] = [[v1]] + [[v2]]
    

    Similarly, View.Map3 takes three existing Views and map their value through a function.

  • View.MapAsync is similar to View.Map but maps through an asynchronous function.

    An important property here is that this combinator saves work by abandoning requests. That is, if the input view changes faster than we can asynchronously convert it, the output view will not propagate change until it obtains a valid latest value. In such a system, intermediate results are thus discarded.

    Similarly, View.MapAsync2 maps two existing Views through an asynchronous function.

  • View.Apply takes a View of a fun