Simplify some sequences that seem unique to rxvt

Almost(?) all other terminals I've tested, shifted numeric keypad keys just
yield the unshifted version of that key. rxvt seems to have sequences for
those keys. Much like Textualize#3737 and Textualize#3739 we *could* add support for these
sequences to the ANSI sequence list, but this is a bit different. Whereas in
those other PRs we've been adding the alternate sequences that rxvt seems
to have for well known and well-supported keys, here it would be the other
way round: it would be adding bindable keys that are only available on a
small subset of environments.

Ideally, at least for the moment, we want to encourage people to bind keys
that are well-supported; so here we turn the shifted keypad keys into their
unshifted counterparts, thus matching many other environments.

See Textualize#3741

Author: davep

src/textual/_ansi_sequences.py

@@ -1,9 +1,11 @@
+from __future__ import annotations
+
 from typing import Mapping, Tuple
 
 from .keys import Keys
 
 # Mapping of vt100 escape codes to Keys.
-ANSI_SEQUENCES_KEYS: Mapping[str, Tuple[Keys, ...]] = {
+ANSI_SEQUENCES_KEYS: Mapping[str, Tuple[Keys, ...] | str] = {
     # Control keys.
     " ": (Keys.Space,),
     "\r": (Keys.Enter,),
@@ -326,6 +328,21 @@
     "\x1b[1;8w": (Keys.Escape, Keys.ControlShift7),
     "\x1b[1;8x": (Keys.Escape, Keys.ControlShift8),
     "\x1b[1;8y": (Keys.Escape, Keys.ControlShift9),
+    # Simplify some sequences that appear to be unique to rxvt; see
+    # https://github.com/Textualize/textual/issues/3741 for context.
+    "\x1bOj": "*",
+    "\x1bOk": "+",
+    "\x1bOm": "-",
+    "\x1bOo": "/",
+    "\x1bOq": "1",
+    "\x1bOr": "2",
+    "\x1bOs": "3",
+    "\x1bOt": "4",
+    "\x1bOu": "5",
+    "\x1bOv": "6",
+    "\x1bOw": "7",
+    "\x1bOx": "8",
+    "\x1bOy": "9",
 }
 
 # https://gist.github.com/christianparpart/d8a62cc1ab659194337d73e399004036

src/textual/_xterm_parser.py

@@ -243,10 +243,20 @@ def _sequence_to_key_events(
             Keys
         """
         keys = ANSI_SEQUENCES_KEYS.get(sequence)
-        if keys is not None:
+        if isinstance(keys, tuple):
+            # If the sequence mapped to a tuple, then it's values from the
+            # `Keys` enum. Raise key events from what we find in the tuple.
             for key in keys:
                 yield events.Key(key.value, sequence if len(sequence) == 1 else None)
-        elif len(sequence) == 1:
+            return
+        # If keys is a string, the intention is that it's a mapping to a
+        # character, which should really be treated as the sequence for the
+        # purposes of the next step...
+        if isinstance(keys, str):
+            sequence = keys
+        # If the sequence is a single character, attempt to process it as a
+        # key.
+        if len(sequence) == 1:
             try:
                 if not sequence.isalnum():
                     name = _character_to_key(sequence)