This page demonstrates how to use FSharp.Formatting.CodeFormat
to tokenize F# source code, obtain information about the source code (mainly tooltips from the type-checker) and how to turn the code into a nicely formatted HTML.
First, we need to load the assembly and open the necessary namespaces:
open FSharp.Formatting.CodeFormat
open System.Reflection
If you want to process multiple snippets, it is a good idea to keep the formatting agent around if possible. The agent needs to load the F# compiler (which needs to load various files itself) and so this takes a long time.
Processing F# sourceThe formatting agent provides a CodeFormatAgent.ParseAndCheckSource method (together with an asynchronous version for use from F# and also a version that returns a .NET Task
for C#). To call the method, we define a simple F# code as a string:
let source =
"""
let hello () =
printfn "Hello world"
"""
let snippets, diagnostics =
CodeFormatter.ParseAndCheckSource("C:\\snippet.fsx", source, None, None, ignore)
When calling the method, you need to specify a file name and the actual content of the script file. The file does not have to physically exist. It is used by the F# compiler to resolve relative references (e.g. #r
) and to automatically name the module including all code in the file.
You can also specify additional parameters, such as *.dll
references, by passing a third argument with compiler options (e.g. "-r:Foo.dll -r:Bar.dll"
).
This operation might take some time, so it is a good idea to use an asynchronous variant of the method. It returns two arrays - the first contains F# snippets in the source code and the second contains any errors reported by the compiler. A single source file can include multiple snippets using the same formatting tags as those used on fssnip.net as documented in the about page.
Working with returned tokensEach returned snippet is essentially just a collection of lines, and each line consists of a sequence of tokens. The following snippet prints basic information about the tokens of our sample snippet:
// Get the first snippet and obtain list of lines
let (Snippet (title, lines)) = snippets |> Seq.head
// Iterate over all lines and all tokens on each line
for (Line (_, tokens)) in lines do
for token in tokens do
match token with
| TokenSpan.Token (kind, code, tip) ->
printf "%s" code
tip
|> Option.iter (fun spans -> printfn "%A" spans)
| TokenSpan.Omitted _
| TokenSpan.Output _
| TokenSpan.Error _ -> ()
printfn ""
The TokenSpan.Token
is the most important kind of token. It consists of a kind (identifier, keyword, etc.), the original F# code and tooltip information. The tooltip is further formatted using a simple document format, but we simply print the value using the F# pretty printing, so the result looks as follows:
let hello[Literal "val hello : unit -> unit"; ...] () =
printfn[Literal "val printfn : TextWriterFormat<'T> -> 'T"; ...] "Hello world"
The Omitted
token is generated if you use the special (*[omit:...]*)
command. The Output
token is generated if you use the // [fsi:...]
command to format output returned by F# interactive. The Error
command wraps code that should be underlined with a red squiggle if the code contains an error.
Finally, the CodeFormat
type also includes a method CodeFormat.FormatHtml that can be used to generate nice HTML output from an F# snippet. This is used, for example, on F# Snippets. The following example shows how to call it:
let prefix = "fst"
let html = CodeFormat.FormatHtml(snippets, prefix)
// Print all snippets, in case there is more of them
for snip in html.Snippets do
printfn "%s" snip.Content
// Print HTML code that is generated for ToolTips
printfn "%s" html.ToolTip
If the input contains multiple snippets separated using the //[snippet:...]
comment, e.g.:
1: 2: 3: 4: 5: 6: 7:
// [snippet: First sample] printf "The answer is: %A" 42 // [/snippet] // [snippet: Second sample] printf "Hello world!" // [/snippet]
then the formatter returns multiple HTML blocks. However, the generated tooltips are shared by all snippets (to save space) and so they are returned separately.
Multiple items
namespace FSharp
--------------------
namespace Microsoft.FSharp
namespace FSharp.Formatting
namespace FSharp.Formatting.CodeFormat
namespace System
namespace System.Reflection
val source: string
val snippets: Snippet array
val diagnostics: SourceError array
module CodeFormatter from FSharp.Formatting.CodeFormat
<summary> Uses agent to handle formatting requests </summary>
val ParseAndCheckSource: file: string * source: string * options: string option * defines: string option * onError: (string -> unit) -> Snippet array * SourceError array
<summary> Parse, check and annotate the source code specified by 'source', assuming that it is located in a specified 'file'. Optional arguments can be used to give compiler command line options and preprocessor definitions </summary>
union case Option.None: Option<'T>
val ignore: value: 'T -> unit
Multiple items
union case Snippet.Snippet: title: string * lines: Line list -> Snippet
--------------------
type Snippet = | Snippet of title: string * lines: Line list
val title: string
val lines: Line list
module Seq from Microsoft.FSharp.Collections
val head: source: 'T seq -> 'T
Multiple items
union case Line.Line: originalLine: string * tokenSpans: TokenSpans -> Line
--------------------
type Line = | Line of originalLine: string * tokenSpans: TokenSpans
val tokens: TokenSpans
val token: TokenSpan
type TokenSpan = | Token of kind: TokenKind * body: string * tip: ToolTipSpans option | Error of kind: ErrorKind * message: string * body: TokenSpans | Omitted of body: string * hidden: string | Output of string
<summary> A token in a parsed F# code snippet. Aside from standard tokens reported from the compiler (Token), this also includes Error (wrapping the underlined tokens), Omitted for the special [omit:...] tags and Output for the special [output:...] tag </summary>
union case TokenSpan.Token: kind: TokenKind * body: string * tip: ToolTipSpans option -> TokenSpan
val kind: TokenKind
val code: string
val tip: ToolTipSpans option
val printf: format: Printf.TextWriterFormat<'T> -> 'T
module Option from Microsoft.FSharp.Core
val iter: action: ('T -> unit) -> option: 'T option -> unit
val spans: ToolTipSpans
val printfn: format: Printf.TextWriterFormat<'T> -> 'T
union case TokenSpan.Omitted: body: string * hidden: string -> TokenSpan
union case TokenSpan.Output: string -> TokenSpan
union case TokenSpan.Error: kind: ErrorKind * message: string * body: TokenSpans -> TokenSpan
Multiple items
type LiteralAttribute = inherit Attribute new: unit -> LiteralAttribute
--------------------
new: unit -> LiteralAttribute
val prefix: string
val html: FormattedContent
type CodeFormat = static member FormatFsx: snippets: Snippet seq -> FormattedContent static member FormatHtml: snippets: Snippet array * prefix: string * ?openTag: string * ?closeTag: string * ?lineNumbers: bool * ?openLinesTag: string * ?closeLinesTag: string * ?addErrors: bool * ?tokenKindToCss: (TokenKind -> string) -> FormattedContent static member FormatLatex: snippets: Snippet array * ?openTag: string * ?closeTag: string * ?lineNumbers: bool -> FormattedContent
<summary> Exposes functionality of the F# code formatter with a nice interface </summary>
static member CodeFormat.FormatHtml: snippets: Snippet array * prefix: string * ?openTag: string * ?closeTag: string * ?lineNumbers: bool * ?openLinesTag: string * ?closeLinesTag: string * ?addErrors: bool * ?tokenKindToCss: (TokenKind -> string) -> FormattedContent
val snip: FormattedSnippet
property FormattedContent.Snippets: FormattedSnippet array with get
<summary> Returns the processed snippets as an array </summary>
property FormattedSnippet.Content: string with get
<summary> Returns the formatted content code for the snipet </summary>
property FormattedContent.ToolTip: string with get
<summary> Returns string with ToolTip elements for all the snippets </summary>
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4