Skip to content
/ ftml Public

Rust library to convert Wikidot code into HTML.

License

Notifications You must be signed in to change notification settings

scpwiki/ftml

Repository files navigation

ftml

Build status Configuration docs.rs link

Foundation Text Markup Language

(Alternatively, ftml: the markup language)

A Rust library to parse Wikidot text ("Wikitext") into an abstract syntax tree (AST). This aims to be a replacement for the aging Text_Wiki from Wikidot. This is version aims to have a nearly fully compatible parser for common Wikidot, including common malformed constructs. The goal is to utilize a lexer generator, and consume the tokens in a custom parser to handle unusual cases with a lax approach.

In addition to providing the speed and safety benefits of Rust, this also improves maintainability, and allows exposing an AST to consumers for more advanced analysis and transformation.

The lint #![forbid(unsafe_code)] is set, and therefore this crate has only safe code. However dependencies may have unsafe internals.

Available under the terms of the GNU Affero General Public License. See LICENSE.md. This library was originally part of Wikijump at /ftml, before being moved to an independent repository, per WJ-1219. Issues for this project will remain in the WJ Jira project.

Compilation

This library targets the latest stable Rust. At time of writing, that is 1.77.0.

$ cargo build --release

You can use this as a dependency by adding the following to your Cargo.toml:

ftml = "1"

The library has two features:

  • html (enabled by default) — This includes the HTML renderer in the crate.
  • mathml (enabled by default) — This includes latex2mathml, which is used to compile any LaTeX into MathML for inclusion in rendered HTML.

They can be disabled by building without features:

$ cargo check --no-default-features

If you wish to build the WebAssembly target for ftml, use wasm-pack:

$ wasm-pack build -- --no-default-features

This optimizes the final WASM, which can take some time. If you are developing and are only interested in the build passing, you should instead use:

$ wasm-pack build --dev

If for some reason you want to invoke cargo check instead, call cargo check --target wasm32-unknown-unkown.

Testing

$ cargo test

Add -- --nocapture to the end if you want to see test output. You can additionally inspect logging by exposing a log-compatible logger.

Philosophy

See Philosophy.md.

Styling

CSS classes are named consistently, in kebab-case only, with prefixes:

  • Any classes with the wj- prefix are those generated automatically, and not intended for direct use by users. An example would be wj-collapsible-block.
  • Any classes with the wiki- prefix are "premade" classes. These are not necessarily generated automatically, but are instead intended for direct use by users wanting to make use of standard styling. An example would be wiki-note.

Naming

"Foundation Text Markup Language" (ftml) is named for the file extension representing in-universe SCP Foundation formatting as mentioned in Kate McTiriss's Proposal. While the expanded form of the initialism is never explicitly stated, it is clearly implied given the name similarity to HTML.

Syntax

ftml is intended to be compatible with a subset of Wikidot text deemed to be "well-formed". Wikidot's general syntax documentation will be relevant here, but weird constructions or strange features may not be. During the development process, they are analyzed and either explicitly unimplemented, or implemented through more sensible syntax. Additionally, it supports several new features and blocks not present in Wikidot, such as checkboxes, and fixes bugs, such as allowing collapsibles to be nested.

As ftml develops into its own branch of wikitext, pages here will document the syntax separately from Wikidot, with the goal of deprecating Wikidot's documentation entirely.

  • Blocks.md -- Which blocks (e.g. [[div]]) are available in ftml and what options they take.

There are some lesser-used or troublesome features which are implemented in a different, backwards-incompatible way with Wikidot. For instance:

  • [[include]] is split into [[include-messy]] (legacy behavior), and [[include-elements]] (self-contained element insertion).
  • Interwiki links are implemented by prefixing ! in triple-bracket links. So [[[!wp:Amazon.com | Amazon]]] instead of [wp:Amazon.com Amazon].

Usage

There are a couple main exported functions, which correspond to each of the main steps in the wikitext process.

First is include, which substitutes all [[include]] blocks for their replaced page content. This returns the substituted wikitext as a new string, as long as the names of all the pages that were used. It requires an object that implement Includer, which handles the process of retrieving pages and generating missing page messages.

Second is preprocess, which will perform Wikidot's various minor text substitutions.

Third is tokenize, which takes the input string and returns a wrapper type. This can be .into()-ed into a Vec<ExtractedToken<'t>> should you want the token extractions it produced. This is used as the input for parse.

Then, borrowing a slice of said tokens, parse consumes them and produces a SyntaxTree representing the full structure of the parsed wikitext.

Finally, with the syntax tree you render it with whatever Render instance you need at the time. Most likely you want HtmlRender. There is also TextRender for text-only, such as for searching article contents or a "printer-friendly" view.

fn include<'t, I, E>(
    input: &'t str,
    includer: I,
    settings: &WikitextSettings,
) -> Result<(String, Vec<PageRef<'t>>), E>
where
    I: Includer<'t, Error = E>;

fn preprocess(
    text: &mut String,
);

fn tokenize<'t>(
    text: &'t str,
) -> Tokenization<'t>;

fn parse<'r, 't>(
    tokenization: &'r Tokenization<'t>,
) -> ParseResult<SyntaxTree<'t>>;

trait Render {
    type Output;

    fn render(
        &self,
        info: &PageInfo,
        tree: &SyntaxTree,
    ) -> Self::Output;
}

When performing a parse, you will need to first run preprocess(), then run parse() on the fully expanded text:

Consider the lifetimes of each of the artifacts being generated, should you want to store the results in a struct.

// Get an `Includer`.
//
// See trait documentation for what this requires, but
// essentially it is some abstract handle that gets the
// contents of a page to be included.
//
// Two sample includers you could try are `NullIncluder`
// and `DebugIncluder`.
let includer = MyIncluderImpl::new();

// Get our source text
let mut input = "**some** test <<string?>>";

// Substitute page inclusions
let (mut text, included_pages) = ftml::include(input, includer, &settings);

// Perform preprocess substitutions
ftml::preprocess(&log, &mut text);

// Generate token from input text
let tokens = ftml::tokenize(&text);

// Parse the token list to produce an AST.
//
// Note that this produces a `ParseResult<SyntaxTree>`, which records the
// parsing warnings in addition to the final result.
let result = ftml::parse(&tokens, &page_info, &settings);

// Here we extract the tree separately from the warning list.
//
// Now we have the final AST, as well as all the issues that
// occurred during the parsing process.
let (tree, warnings) = result.into();

// Finally, we render with our renderer. Generally this is `HtmlRender`,
// but you could have a custom implementation here too.
//
// You must provide a `PageInfo` struct, which describes the page being rendered.
// You must also provide a handle to provide various remote sources, such as
// module content, but this is not stabilized yet.
let html_output = HtmlRender.render(&tree, &page_info, &settings);

JSON Serialization

See Serialization.md.