[text] add 'sanitize_whitespace()'

gallery_dl/text.py

@@ -66,6 +66,11 @@ def slugify(value):
     return re(r"[-\s]+").sub("-", value).strip("-_")
 
 
+def sanitize_whitespace(value):
+    """Replace all whitespace characters with a single space"""
+    return re(r"\s+").sub(" ", value.strip())
+
+
 def ensure_http_scheme(url, scheme="https://"):
     """Prepend 'scheme' to 'url' if it doesn't have one"""
     if url and not url.startswith(("https://", "http://")):

test/test_text.py

@@ -106,6 +106,17 @@ class TestText(unittest.TestCase):
         self.assertEqual(f(1), "1")
         self.assertEqual(f(2.3), "23")
 
+    def test_sanitize_whitespace(self, f=text.sanitize_whitespace):
+        self.assertEqual(f("Hello World"), "Hello World")
+        self.assertEqual(f("Hello\tWorld"), "Hello World")
+        self.assertEqual(f("  Hello   World  "), "Hello World")
+        self.assertEqual(f("\tHello  \n\tWorld  "), "Hello World")
+
+        self.assertEqual(f(""), "")
+        self.assertEqual(f(" "), "")
+        self.assertEqual(f("      "), "")
+        self.assertEqual(f(" \t\n "), "")
+
     def test_ensure_http_scheme(self, f=text.ensure_http_scheme):
         result = "https://example.org/filename.ext"