Add remove_multiple_whitespaces function to remove multiple whitespac…

…es from text

mltb2/text.py

@@ -4,7 +4,8 @@
 
 """Text specific module."""
 
-from typing import Dict, Final, Tuple
+import re
+from typing import Dict, Final, Pattern, Tuple
 
 INVISIBLE_CHARACTERS: Final[Tuple[str, ...]] = (
     "\u200b",  # Zero Width Space (ZWSP) https://www.compart.com/en/unicode/U+200b
@@ -37,6 +38,9 @@
 SPECIAL_WHITESPACES_TRANS: Final[Dict[int, str]] = str.maketrans({char: " " for char in SPECIAL_WHITESPACES})
 
 
+MULTI_SPACE_PATTERN: Pattern = re.compile(r" {2,}")
+
+
 def remove_invisible_characters(text: str) -> str:
     """Remove invisible characters from text.
 
@@ -91,3 +95,15 @@ def has_special_whitespaces(text: str) -> bool:
         ``True`` if the text contains special whitespaces, ``False`` otherwise.
     """
     return any(char in text for char in SPECIAL_WHITESPACES)
+
+
+def remove_multiple_whitespaces(text: str) -> str:
+    """Remove multiple whitespaces from text.
+
+    Args:
+        text: The text from which the multiple whitespaces are to be removed.
+
+    Returns:
+        The cleaned text.
+    """
+    return MULTI_SPACE_PATTERN.sub(" ", text)

tests/test_text.py

@@ -10,6 +10,7 @@
     has_invisible_characters,
     has_special_whitespaces,
     remove_invisible_characters,
+    remove_multiple_whitespaces,
     replace_special_whitespaces,
 )
 
@@ -74,3 +75,27 @@ def test_has_special_whitespaces_false():
     text = "Hello you!"
     result = has_special_whitespaces(text)
     assert not result
+
+
+def test_remove_multiple_whitespaces():
+    text = "Hello  World  !"
+    result = remove_multiple_whitespaces(text)
+    assert result == "Hello World !"
+
+
+def test_remove_multiple_whitespaces_empty():
+    text = ""
+    result = remove_multiple_whitespaces(text)
+    assert result == ""
+
+
+def test_remove_multiple_whitespaces_empty_result():
+    text = "  "
+    result = remove_multiple_whitespaces(text)
+    assert result == " "
+
+
+def test_remove_multiple_whitespaces_one_space():
+    text = " "
+    result = remove_multiple_whitespaces(text)
+    assert result == " "