Allow conversion to and from unicode names.

Throughout the code base, there are a couple of places (noted in #1830) where we take a Unicode name and make it more friendly, or take a 'friendly' name and try to recover the original Unicode name. We add a function to go from 'friendly' to the Unicode name, which tackles the issue highlighted in #1815.

Related issues: #1815, #1830.

src/textual/keys.py

@@ -209,6 +209,37 @@ class Keys(str, Enum):
 }
 REPLACED_KEYS = {value: key for key, value in KEY_NAME_REPLACEMENTS.items()}
 
+# Convert the friendly versions of character key Unicode names
+# back to their original names.
+# This is because we go from Unicode to friendly by replacing spaces and dashes
+# with underscores, which cannot be undone by replacing underscores with spaces/dashes.
+KEY_TO_UNICODE_NAME = {
+    "exclamation_mark": "EXCLAMATION MARK",
+    "quotation_mark": "QUOTATION MARK",
+    "number_sign": "NUMBER SIGN",
+    "dollar_sign": "DOLLAR SIGN",
+    "percent_sign": "PERCENT SIGN",
+    "left_parenthesis": "LEFT PARENTHESIS",
+    "right_parenthesis": "RIGHT PARENTHESIS",
+    "plus_sign": "PLUS SIGN",
+    "hyphen_minus": "HYPHEN-MINUS",
+    "full_stop": "FULL STOP",
+    "less_than_sign": "LESS-THAN SIGN",
+    "equals_sign": "EQUALS SIGN",
+    "greater_than_sign": "GREATER-THAN SIGN",
+    "question_mark": "QUESTION MARK",
+    "commercial_at": "COMMERCIAL AT",
+    "left_square_bracket": "LEFT SQUARE BRACKET",
+    "reverse_solidus": "REVERSE SOLIDUS",
+    "right_square_bracket": "RIGHT SQUARE BRACKET",
+    "circumflex_accent": "CIRCUMFLEX ACCENT",
+    "low_line": "LOW LINE",
+    "grave_accent": "GRAVE ACCENT",
+    "left_curly_bracket": "LEFT CURLY BRACKET",
+    "vertical_line": "VERTICAL LINE",
+    "right_curly_bracket": "RIGHT CURLY BRACKET",
+}
+
 # Some keys have aliases. For example, if you press `ctrl+m` on your keyboard,
 # it's treated the same way as if you press `enter`. Key handlers `key_ctrl_m` and
 # `key_enter` are both valid in this case.
@@ -233,6 +264,14 @@ class Keys(str, Enum):
 }
 
 
+def _get_unicode_name_from_key(key: str) -> str:
+    """Get the best guess for the Unicode name of the char corresponding to the key.
+
+    This function can be seen as a pseudo-inverse of the function `_character_to_key`.
+    """
+    return KEY_TO_UNICODE_NAME.get(key, key.upper())
+
+
 def _get_key_aliases(key: str) -> list[str]:
     """Return all aliases for the given key, including the key itself"""
     return [key] + KEY_ALIASES.get(key, [])
@@ -261,8 +300,10 @@ def _get_key_display(key: str) -> str:
 
 
 def _character_to_key(character: str) -> str:
-    """Convert a single character to a key value."""
-    assert len(character) == 1
+    """Convert a single character to a key value.
+
+    This transformation can be undone by the function `_get_unicode_name_from_key`.
+    """
     if not character.isalnum():
         key = unicodedata.name(character).lower().replace("-", "_").replace(" ", "_")
     else: