structform

Crates.iostructform
lib.rsstructform
version0.1.0
sourcesrc
created_at2021-02-11 14:27:54.906135
updated_at2021-02-11 14:27:54.906135
descriptionLibrary for managing interactive forms which encodes validation rules using the type system.
homepage
repositoryhttps://github.com/panoptix-za/structform
max_upload_size
id353786
size68,465
OSS maintainers (github:panoptix-za:oss-maintainers)

documentation

README

StructForm

crates.io crates.io build status

Library for managing interactive forms which encodes validation rules using the type system.

Why StructForm

Forms are a common way of capturing data in many programs. It's nice in these programs if the form can give realtime feedback on the data being captured. Additionally, if you have a separate frontend and backend communicating over an API, it's good to keep the validation rules enforced by the frontend and backend the same.

Rust has an excellent type system. StructForm is designed to leverage the type system to give realtime feedback and enforce validation rules.

StructForm was designed with web applications that follow the Elm architecture in mind (specifically we use Seed), but it should work well with any frontend framework.

How Does it Work?

It all starts with a strongly typed version of the data you want your form to represent.

#[derive(Default, Debug, PartialEq, Eq)]
struct LoginData {
    username: String,
    password: String,
}

This is mirrorred by a Form struct, that derives the StructForm trait.

#[derive(Default, Clone, StructForm)]
#[structform(model = "LoginData")]
struct LoginForm {
    username: FormTextInput<String>,
    password: FormPasswordInput<String>,
}

This will do a few things:

  • An enum will be created, named LoginFormField in this example, which has a case for each of your inputs.
  • A set_input function is implemented on your form, which takes the field enum and a string and updates the appropriate field in your form.
  • A submit function is implemented on your form, which checks all of the inputs on your form and, if possible, creates your strongly typed model (LoginData in this example).

In a typical web form, this is how you would hook it up:

  • Have an HTML input for each field in your form.
  • Add a listener to the input event on the inputs to call your form's set_input. Conveniently, all of your inputs can use the same event callback as long as they pass through the appropriate value from the form field enum.
  • Add a listener to the submit event on your HTML form, which calls submit on your form and, if submit is successful, send your data on to an API.

The best way to learn to use StructForm is to look at the examples.

Form Inputs

StructForm is unusual in that it doesn't come with form inputs. Rather, it gives you macros that you use to derive your own form inputs. This lets you specify which types of inputs you have in your application, and how those different inputs might convert from strings typed by your end users to strongly typed data differently.

The basic way of creating a form input is to call the derive_form_input macro.

derive_form_input! {MyFormInput}

This will create a struct for you called MyFormInput, with an input string that you tie to your UI. You can now specify how MyFormInput handles any model type, including details like trimming the input strings or not.

impl ParseAndFormat<MyInputModel> for MyFormInput<Model> {
    fn parse(value: &str) -> Result<MyInputModel, ParseError> { todo!() }
    fn format(value: &MyInputModel) -> String { todo!() }
}

For some common implementations of ParseAndFormat, see the macros impl_text_input_with_stringops and impl_numeric_input_with_stringops.

Validation

Validation should be added by making the types that your form inputs wrap enforce your validation rules. The newtype pattern is an excellent way to add these rules. See the validation rules example for an example of how this can be done.

Examples

License

Licensed under either of

at your option.

Commit count: 15

cargo fmt