glossa-DSL

old_name: tmpl_resolver

A domain-specific language designed exclusively for localization (L10n).

Its syntax is similar to Mozilla Fluent, but it doesn't have as many features as Fluent.

The core logic is implemented using nom and can be used in no_std.

Key Concepts

• Templates: Contain either direct text parts or conditional selectors
• Selectors: Enable branch logic based on parameter values
• Variable Resolution: Recursive resolution with context-aware lookup

Features

• []
  • Minimal configuration for no_std use
• ["std"]
  • Enables standard library
  • Uses ahash::HashMap for faster lookups
• ["serde"]
  • Adds serialization capabilities
  • Enables template storage/transmission
• ["bincode"]
  • Efficient binary serialization
• ["toml"]
  • Enables ResolverError::{DecodeTomlError, EncodeTomlError}

Basic

use glossa_dsl::{Resolver, error::ResolverResult};

fn main() -> ResolverResult<()> {
  let resolver: Resolver = [
      ("h", "Hello"),
      ("greeting", "{h} { $name }! Today is {$day}.")
    ]
    .try_into()?;

  let ctx = [("name", "Alice"), ("day", "Sunday")];

  let result = resolver.get_with_context("greeting", &ctx)?;
  assert_eq!(result, "Hello Alice! Today is Sunday.");
  Ok(())
}

Conditional Logic

use glossa_dsl::{Resolver, error::ResolverResult};

fn main() -> ResolverResult<()> {
  let selector_msg = [(
    "message",
    r#"
    $status ->
      [success] Operation succeeded!
      [error] Error occurred!
      *[default] Unknown status: {$status}
    "#
  )];

  let resolver: Resolver = selector_msg.try_into()?;

  let success_msg = resolver.get_with_context("message", &[("status", "success")])?;

  assert_eq!(success_msg, "Operation succeeded!");
  Ok(())
}

Escape

• "{{ a }}" => "a"
• "{{{a}}}" => "a"
• "{{{{ a }}}}" => "a"
• "{{ {a} }}" => "{a}"
• "{{a}" => ❌ nom Error, code: take_until
• "{{{ {{a}} }}}" => "{{a}}"
• "{{{ {{ a }} }}}" => "{{ a }}"
• "{{{ {{a} }}}" => "{{a}"

use glossa_dsl::{error::ResolverResult, Resolver};

fn main() -> ResolverResult<()> {
  let resolver: Resolver = [
    ("h", "Hello { $name }"),
    ("how_are_you", "How Are You"),
    ("greeting", "{h}!{{ how_are_you }}? {{    {$name} }}"),
  ]
  .try_into()?;

  let ctx = [("name", "Alice")];

  let result = resolver.get_with_context("greeting", &ctx)?;
  assert_eq!(result, "Hello Alice!how_are_you? {$name}");
  Ok(())
}

Real World Examples

Add dependencies

cargo add toml tap anyhow
cargo add glossa-dsl --features=std,serde,toml

Emoji

We can use emoji as variable identifier name.

toml:

"🐱" = "ฅ(°ω°ฅ)"
hello = "Hello {🐱}"

1. hello references {🐱}
2. expanding hello
3. we would get "Hello ฅ(°ω°ฅ)".

rust:

let text = res.try_get("hello")?;
assert_eq!(text, "Hello ฅ(°ω°ฅ)");

toml:

hello = "Hello {$🐱}"

$🐱 means that its value is passed in externally.

rust:

let text = res.get_with_context("hello", &[("🐱", "QwQ")])?;
assert_eq!(text, "Hello QwQ");

use tap::Pipe;
use glossa_dsl::{error::ResolverResult, Resolver, resolver::AHashRawMap};

fn main() -> ResolverResult<()> {
  let res: Resolver = r##"
      meow = "喵"
      "🐱" = "{ meow } ฅ(°ω°ฅ)"

      "问候" = """
        $period ->
          [morning] 早安{🐱}
          [night] 晚安{🐱}
          *[other] {$period}好
      """

      "称谓" = """
      $gender ->
        *[male] 先生
        [female] 女士
      """

      greeting = "{ 问候 }！{ $name }{ 称谓 }。"
    "##
    .pipe(toml::from_str::<AHashRawMap>)?
    .try_into()?;

  let get_text = |ctx| res.get_with_context("greeting", ctx);

  let text = [
    ("period", "morning"),
    ("name", "Young"),
    ("gender", "unknown"),
  ]
  .as_ref()
  .pipe(get_text)?;

  assert_eq!(text, "早安喵 ฅ(°ω°ฅ)！Young先生。");
  assert_eq!(res.try_get("🐱")?, "喵 ฅ(°ω°ฅ)");

  Ok(())
}

Simple L10n Message

use anyhow::Result as AnyResult;
use tap::{Pipe, TryConv};
use glossa_dsl::{Resolver, resolver::AHashRawMap};

const EN_TOML: &str = r#"
  num-to-en = """
    $num ->
      [0] zero
      [1] one
      [2] two
      [3] three
      *[other] {$num}
  """

  unread_msg = "unread message"

  unread-count = """
    $num ->
      [0] No {unread_msg}s.
      [1] You have { num-to-en } {unread_msg}.
      *[other] You have { num-to-en } {unread_msg}s.
  """

  show-unread-messages-count = "{unread-count}"
"#;

const ZH_TOML: &str = r#"
  "阿拉伯数字转汉字" = """
    $num ->
      [0] 〇
      [1] 一
      [2] 二
      [3] 三
      *[其他] {$num}
  """

  "未读msg" = "未读消息"

  "显示未读消息数量" = """
    $num ->
        [0] 没有{ 未读msg }。
        [2] 您有两条{ 未读msg }。
       *[其他] 您有{ 阿拉伯数字转汉字 }条{ 未读msg }。
  """

  show-unread-messages-count = "{显示未读消息数量}"
"#;

fn main() -> AnyResult<()> {
  let get_text = |lang| -> AnyResult<_> {
    match lang {
      "zh" => ZH_TOML,
      _ => EN_TOML,
    }
    .pipe(toml::from_str::<AHashRawMap>)?
    .try_conv::<Resolver>()?
    .pipe(|r| {
      move |num_str| {
        r.get_with_context("show-unread-messages-count", &[("num", num_str)])
      }
    })
    .pipe(Ok)
  };

  let get_en_text = get_text("en")?;
  assert_eq!(get_en_text("0")?, "No unread messages.");
  assert_eq!(get_en_text("1")?, "You have one unread message.");
  assert_eq!(get_en_text("2")?, "You have two unread messages.");
  assert_eq!(get_en_text("100")?, "You have 100 unread messages.");

  let get_zh_text = get_text("zh")?;
  assert_eq!(get_zh_text("0")?, "没有未读消息。");
  assert_eq!(get_zh_text("1")?, "您有一条未读消息。");
  assert_eq!(get_zh_text("2")?, "您有两条未读消息。");
  assert_eq!(get_zh_text("100")?, "您有100条未读消息。");

  Ok(())
}