feat: Smarter default language detection (#26)

README-ja.md

@@ -91,8 +91,8 @@ js-i18n.nvim は、JavaScript の i18n ライブラリをサポートする Neov
 ```lua
 {
   primary_language = {}, -- 優先表示する言語（バーチャルテキストなどの表示に使用する言語の初期設定）
-  translation_source = { "**/locales/*/translation.json" }, -- 翻訳リソースのパターン
-  detect_language = ..., -- 言語を検出する関数。
+  translation_source = { "**/{locales,messages}/*.json" }, -- 翻訳リソースのパターン
+  detect_language = ..., -- 言語を検出する関数。デフォルトではファイル名からヒューリスティックに検出する関数が使用されます。
   key_separator = ".", -- キーのセパレータ
   virt_text = {
     enabled = true, -- バーチャルテキストの表示を有効にする
@@ -113,4 +113,3 @@ js-i18n.nvim は、JavaScript の i18n ライブラリをサポートする Neov
 
 - ライブラリサポートの強化
   - namespace のサポート
-- 今より賢い翻訳リソースの検出

README.md

@@ -96,8 +96,8 @@ The default settings are as follows. For omitted parts, refer to [config.lua](./
 ```lua
 {
   primary_language = {}, -- The default language to display (initial setting for displaying virtual text, etc.)
-  translation_source = { "**/locales/*/translation.json" }, -- Pattern for translation resources
-  detect_language = ..., -- Function to detect the language.
+  translation_source = { "**/{locales,messages}/*.json" }, -- Pattern for translation resources
+  detect_language = ..., -- Function to detect the language. By default, a function that detects the language heuristically from the file name is used.
   key_separator = ".", -- Key separator
   virt_text = {
     enabled = true, -- Enable virtual text display
@@ -118,4 +118,3 @@ The default settings are as follows. For omitted parts, refer to [config.lua](./
 
 - Enhanced support for libraries
   - Namespace support
-- Smarter translation resource detection

lua/js-i18n/config.lua

@@ -1,13 +1,32 @@
+local langs = require("js-i18n.lang_name_list")
 local utils = require("js-i18n.utils")
 
 local M = {}
 
---- ファイルパスから言語を検出する
---- @param path string ファイルパス
-local function default_detect_language(path)
+local function normalize_lang(locale)
+  return locale:lower():gsub("-", "_")
+end
+
+local LangSet = {}
+for _, l in ipairs(langs) do
+  LangSet[normalize_lang(l)] = true
+end
+
+--- Detect language from file path heuristically
+--- @param path string File path
+function M.default_detect_language(path)
   local abs_path = vim.fn.fnamemodify(path, ":p")
-  local split = vim.split(abs_path, "/")
-  local lang = split[#split - 1]
+  local split = vim.split(abs_path, "[/.]")
+
+  local lang = nil
+
+  for _, part in ipairs(vim.fn.reverse(split)) do
+    if LangSet[normalize_lang(part)] then
+      lang = part
+      break
+    end
+  end
+
   return lang
 end
 
@@ -56,8 +75,8 @@ end
 --- @type I18n.Config
 local default_config = {
   primary_language = {},
-  translation_source = { "**/locales/**/*.json" },
-  detect_language = default_detect_language,
+  translation_source = { "**/{locales,messages}/*.json" },
+  detect_language = M.default_detect_language,
   key_separator = ".",
   virt_text = {
     enabled = true,

lua/js-i18n/lang_name_list.lua

@@ -0,0 +1,236 @@
+-- http://tools.ietf.org/html/rfc5646
+
+return {
+  "af",
+  "af-ZA",
+  "ar",
+  "ar-AE",
+  "ar-BH",
+  "ar-DZ",
+  "ar-EG",
+  "ar-IQ",
+  "ar-JO",
+  "ar-KW",
+  "ar-LB",
+  "ar-LY",
+  "ar-MA",
+  "ar-OM",
+  "ar-QA",
+  "ar-SA",
+  "ar-SY",
+  "ar-TN",
+  "ar-YE",
+  "az",
+  "az-AZ",
+  "az-Cyrl-AZ",
+  "be",
+  "be-BY",
+  "bg",
+  "bg-BG",
+  "bs-BA",
+  "ca",
+  "ca-ES",
+  "cs",
+  "cs-CZ",
+  "cy",
+  "cy-GB",
+  "da",
+  "da-DK",
+  "de",
+  "de-AT",
+  "de-CH",
+  "de-DE",
+  "de-LI",
+  "de-LU",
+  "dv",
+  "dv-MV",
+  "el",
+  "el-GR",
+  "en",
+  "en-AU",
+  "en-BZ",
+  "en-CA",
+  "en-CB",
+  "en-GB",
+  "en-IE",
+  "en-JM",
+  "en-NZ",
+  "en-PH",
+  "en-TT",
+  "en-US",
+  "en-ZA",
+  "en-ZW",
+  "eo",
+  "es",
+  "es-AR",
+  "es-BO",
+  "es-CL",
+  "es-CO",
+  "es-CR",
+  "es-DO",
+  "es-EC",
+  "es-ES",
+  "es-GT",
+  "es-HN",
+  "es-MX",
+  "es-NI",
+  "es-PA",
+  "es-PE",
+  "es-PR",
+  "es-PY",
+  "es-SV",
+  "es-UY",
+  "es-VE",
+  "et",
+  "et-EE",
+  "eu",
+  "eu-ES",
+  "fa",
+  "fa-IR",
+  "fi",
+  "fi-FI",
+  "fo",
+  "fo-FO",
+  "fr",
+  "fr-BE",
+  "fr-CA",
+  "fr-CH",
+  "fr-FR",
+  "fr-LU",
+  "fr-MC",
+  "gl",
+  "gl-ES",
+  "gu",
+  "gu-IN",
+  "he",
+  "he-IL",
+  "hi",
+  "hi-IN",
+  "hr",
+  "hr-BA",
+  "hr-HR",
+  "hu",
+  "hu-HU",
+  "hy",
+  "hy-AM",
+  "id",
+  "id-ID",
+  "is",
+  "is-IS",
+  "it",
+  "it-CH",
+  "it-IT",
+  "ja",
+  "ja-JP",
+  "ka",
+  "ka-GE",
+  "kk",
+  "kk-KZ",
+  "kn",
+  "kn-IN",
+  "ko",
+  "ko-KR",
+  "kok",
+  "kok-IN",
+  "ky",
+  "ky-KG",
+  "lt",
+  "lt-LT",
+  "lv",
+  "lv-LV",
+  "mi",
+  "mi-NZ",
+  "mk",
+  "mk-MK",
+  "mn",
+  "mn-MN",
+  "mr",
+  "mr-IN",
+  "ms",
+  "ms-BN",
+  "ms-MY",
+  "mt",
+  "mt-MT",
+  "nb",
+  "nb-NO",
+  "nl",
+  "nl-BE",
+  "nl-NL",
+  "nn-NO",
+  "ns",
+  "ns-ZA",
+  "pa",
+  "pa-IN",
+  "pl",
+  "pl-PL",
+  "ps",
+  "ps-AR",
+  "pt",
+  "pt-BR",
+  "pt-PT",
+  "qu",
+  "qu-BO",
+  "qu-EC",
+  "qu-PE",
+  "ro",
+  "ro-RO",
+  "ru",
+  "ru-RU",
+  "sa",
+  "sa-IN",
+  "se",
+  "se-FI",
+  "se-NO",
+  "se-SE",
+  "sk",
+  "sk-SK",
+  "sl",
+  "sl-SI",
+  "sq",
+  "sq-AL",
+  "sr-BA",
+  "sr-Cyrl-BA",
+  "sr-SP",
+  "sr-Cyrl-SP",
+  "sv",
+  "sv-FI",
+  "sv-SE",
+  "sw",
+  "sw-KE",
+  "syr",
+  "syr-SY",
+  "ta",
+  "ta-IN",
+  "te",
+  "te-IN",
+  "th",
+  "th-TH",
+  "tl",
+  "tl-PH",
+  "tn",
+  "tn-ZA",
+  "tr",
+  "tr-TR",
+  "tt",
+  "tt-RU",
+  "ts",
+  "uk",
+  "uk-UA",
+  "ur",
+  "ur-PK",
+  "uz",
+  "uz-UZ",
+  "uz-Cyrl-UZ",
+  "vi",
+  "vi-VN",
+  "xh",
+  "xh-ZA",
+  "zh",
+  "zh-CN",
+  "zh-HK",
+  "zh-MO",
+  "zh-SG",
+  "zh-TW",
+  "zu",
+  "zu-ZA",
+}

tests/js-i18n/config_spec.lua

@@ -0,0 +1,33 @@
+local config = require("js-i18n.config")
+
+describe("js-i18n.config", function()
+  describe("default_detect_language", function()
+    local tests = {
+      { path = "/path/to/locals/en/trans.json", expected = "en" },
+      { path = "/path/to/locals/ja/trans.json", expected = "ja" },
+      { path = "/path/to/locals/hoge/trans.json", expected = nil },
+
+      -- Test cases to verify that it is sufficient for the languagee name to be included somewhere.
+      { path = "/path/to/locals/sub/en.json", expected = "en" },
+      { path = "/path/to/en/locals/trans.json", expected = "en" },
+      { path = "/path/to/locals/en-trans.json", expected = nil },
+
+      -- Test cases for language names with any case and separating characters.
+      { path = "/path/to/locals/en-us/trans.json", expected = "en-us" },
+      { path = "/path/to/locals/en_us/trans.json", expected = "en_us" },
+      { path = "/path/to/locals/en-US/trans.json", expected = "en-US" },
+
+      -- Test cases where the last match is returned when multiple locale names are included.
+      { path = "/path/to/locals/en/ja.json", expected = "ja" },
+    }
+
+    for _, test in ipairs(tests) do
+      it(
+        string.format("should return %q when detecting language from %q", test.expected, test.path),
+        function()
+          assert.are.equal(test.expected, config.default_detect_language(test.path))
+        end
+      )
+    end
+  end)
+end)