feat: can find words that include special characters

Previously, only words that consisted of letters, numbers, underscores,
and dashes were matched. This change allows for words that include
special word like characters such as

- umlauts (e.g. 'ö', 'ä', 'å', 'ü')
- accented characters (e.g. 'é', 'ç', 'ñ', 'ü')
- Cyrillic characters (e.g. 'к', 'о', 'с', 'м')
- Chinese characters (e.g. '你', '好')
- Japanese characters (e.g. '日', '本', '語')
- Korean characters (e.g. '한', '국', '어')
- Greek characters (e.g. 'τ', 'ο', 'π', 'ι', 'κ', 'ή')
- emoji (e.g. '😄', '👍', '👎')

lua/blink-ripgrep/init.lua

@@ -13,12 +13,34 @@
 ---@field get_completions? fun(self: blink.cmp.Source, context: blink.cmp.Context, callback: fun(response: blink.cmp.CompletionResponse | nil)):  nil
 local RgSource = {}
 
+local word_character = vim.lpeg.R("az", "AZ", "09", "\128\255")
+  + vim.lpeg.P("_")
+  + vim.lpeg.P("-")
+local non_word_character = vim.lpeg.P(1) - word_character
+
+local collect_pattern = vim.lpeg.Ct(
+  -- Skip non-word characters first
+  (
+    non_word_character ^ 0
+    * vim.lpeg.C(word_character ^ 1)
+    * non_word_character ^ 0
+  ) ^ 0
+)
+
+---@param text_before_cursor string "The text of the entire line before the cursor"
+---@return string
+function RgSource.match_prefix(text_before_cursor)
+  local matches = vim.lpeg.match(collect_pattern, text_before_cursor)
+  return matches and matches[#matches] or ""
+end
+
 ---@param context blink.cmp.Context
 ---@return string
 local function default_get_prefix(context)
   local line = context.line
   local col = context.cursor[2]
-  local prefix = line:sub(1, col):match("[%w_-]+$") or ""
+  local text = line:sub(1, col)
+  local prefix = RgSource.match_prefix(text)
   return prefix
 end
 

spec/blink-ripgrep/get_prefix_spec.lua

@@ -0,0 +1,129 @@
+local assert = require("luassert")
+local blink_ripgrep = require("blink-ripgrep")
+
+describe("match_prefix", function()
+  it("for simple strings", function()
+    assert.are_same(blink_ripgrep.match_prefix("hello"), "hello")
+    assert.are_same(blink_ripgrep.match_prefix("abc123"), "abc123")
+    assert.are_same(blink_ripgrep.match_prefix("abc-123"), "abc-123")
+    assert.are_same(blink_ripgrep.match_prefix("abc_123"), "abc_123")
+  end)
+
+  it(
+    "matches when there is one nonmatching piece of input in the beginning",
+    function()
+      assert.are_same(blink_ripgrep.match_prefix(".hello"), "hello")
+      assert.are_same(blink_ripgrep.match_prefix(",hello"), "hello")
+      assert.are_same(
+        blink_ripgrep.match_prefix(".,,!@!@$@%<<@$<hello"),
+        "hello"
+      )
+      assert.are_same(blink_ripgrep.match_prefix(" hello"), "hello")
+      assert.are_same(blink_ripgrep.match_prefix("random_text hello"), "hello")
+      assert.are_same(blink_ripgrep.match_prefix("-- hello"), "hello")
+      assert.are_same(blink_ripgrep.match_prefix("-- abc123"), "abc123")
+      assert.are_same(blink_ripgrep.match_prefix("-- abc-123"), "abc-123")
+      assert.are_same(blink_ripgrep.match_prefix("-- abc_123"), "abc_123")
+    end
+  )
+
+  it(
+    "matches when there are multiple nonmatching pieces of input in the beginning",
+    function()
+      assert.are_same(blink_ripgrep.match_prefix(".hello.hello"), "hello")
+      assert.are_same(blink_ripgrep.match_prefix(",hello,hello"), "hello")
+      assert.are_same(
+        blink_ripgrep.match_prefix(".,,!@!@$@%<<@$<hello.,,!@!@$@%<<@$<hello"),
+        "hello"
+      )
+      assert.are_same(blink_ripgrep.match_prefix(" hello hello"), "hello")
+
+      assert.are_same(
+        blink_ripgrep.match_prefix("random_text hello hello"),
+        "hello"
+      )
+      assert.are_same(blink_ripgrep.match_prefix("-- hello hello"), "hello")
+      assert.are_same(blink_ripgrep.match_prefix("-- abc123 abc123"), "abc123")
+      assert.are_same(
+        blink_ripgrep.match_prefix("-- abc-123 abc-123"),
+        "abc-123"
+      )
+      assert.are_same(
+        blink_ripgrep.match_prefix("-- abc_123 abc_123"),
+        "abc_123"
+      )
+    end
+  )
+
+  it("for multipart strings", function()
+    -- three parts
+    assert.are_same(
+      blink_ripgrep.match_prefix("hello-world-today"),
+      "hello-world-today"
+    )
+
+    -- three parts with numbers
+    assert.are_same(
+      blink_ripgrep.match_prefix("abc123-def456-ghi789"),
+      "abc123-def456-ghi789"
+    )
+
+    -- multiple parts with mixed dashes and underscores
+    assert.are_same(
+      blink_ripgrep.match_prefix("abc-123_def-456_ghi-789"),
+      "abc-123_def-456_ghi-789"
+    )
+  end)
+
+  it("matches special characters", function()
+    -- umlauts and other special characters
+    assert.are_same(blink_ripgrep.match_prefix("yöllä"), "yöllä") -- Finnish word with 'ö' and 'ä'
+    assert.are_same(blink_ripgrep.match_prefix("über"), "über") -- German word with 'ü'
+    assert.are_same(blink_ripgrep.match_prefix("übermensch"), "übermensch") -- German compound word with 'ü'
+    assert.are_same(blink_ripgrep.match_prefix("mañana"), "mañana") -- Spanish word with 'ñ'
+    assert.are_same(blink_ripgrep.match_prefix("Ångström"), "Ångström") -- Swedish word with 'Å' and 'ö'
+    assert.are_same(blink_ripgrep.match_prefix("Straße"), "Straße") -- German word with 'ß'
+    assert.are_same(blink_ripgrep.match_prefix("český"), "český") -- Czech word with 'č'
+    assert.are_same(blink_ripgrep.match_prefix("naïve"), "naïve") -- French word with 'ï'
+    assert.are_same(blink_ripgrep.match_prefix("façade"), "façade") -- French word with 'ç'
+    assert.are_same(blink_ripgrep.match_prefix("résumé"), "résumé") -- French word with 'é'
+    assert.are_same(blink_ripgrep.match_prefix("космос"), "космос") -- Russian word with Cyrillic characters
+    assert.are_same(blink_ripgrep.match_prefix("你好"), "你好") -- Chinese characters
+    assert.are_same(blink_ripgrep.match_prefix("日本語"), "日本語") -- Japanese characters
+    assert.are_same(blink_ripgrep.match_prefix("한국어"), "한국어") -- Korean characters
+    assert.are_same(blink_ripgrep.match_prefix("τοπική"), "τοπική") -- Greek word with 'π' and 'ή'
+  end)
+
+  it("matches emoji", function()
+    -- because why not 😄
+    assert.are_same(blink_ripgrep.match_prefix("👍👎"), "👍👎")
+    assert.are_same(blink_ripgrep.match_prefix("👍-👎"), "👍-👎")
+  end)
+
+  it("does not include punctuation characters", function()
+    assert.are_same(blink_ripgrep.match_prefix("!hello"), "hello")
+    assert.are_same(blink_ripgrep.match_prefix("?world"), "world")
+    assert.are_same(blink_ripgrep.match_prefix("#hashtag"), "hashtag")
+    assert.are_same(blink_ripgrep.match_prefix("$money"), "money")
+    assert.are_same(blink_ripgrep.match_prefix("%value"), "value")
+    assert.are_same(blink_ripgrep.match_prefix("&and"), "and")
+    assert.are_same(blink_ripgrep.match_prefix("*star"), "star")
+    assert.are_same(blink_ripgrep.match_prefix("@email"), "email")
+    assert.are_same(blink_ripgrep.match_prefix("~tilde"), "tilde")
+    assert.are_same(blink_ripgrep.match_prefix(";semicolon"), "semicolon")
+    assert.are_same(blink_ripgrep.match_prefix(":colon"), "colon")
+  end)
+
+  it("does not include whitespace and control characters", function()
+    assert.are_same(blink_ripgrep.match_prefix(" hello"), "hello")
+    assert.are_same(blink_ripgrep.match_prefix("world "), "world")
+    assert.are_same(blink_ripgrep.match_prefix("\t\ttext"), "text")
+    assert.are_same(blink_ripgrep.match_prefix("\nnewline"), "newline")
+  end)
+
+  it("includes symbols", function()
+    assert.are_same(blink_ripgrep.match_prefix("©copyright"), "©copyright")
+    assert.are_same(blink_ripgrep.match_prefix("®registered"), "®registered")
+    assert.are_same(blink_ripgrep.match_prefix("™trademark"), "™trademark")
+  end)
+end)