Include radix base prefix in large number representation (#7700)

## Summary

When lexing a number like `0x995DC9BBDF1939FA` that exceeds our small
number representation, we were only storing the portion after the base
(in this case, `995DC9BBDF1939FA`). When using that representation in
code generation, this could lead to invalid syntax, since
`995DC9BBDF1939FA)` on its own is not a valid integer.

This PR modifies the code to store the full span, including the radix
prefix.

See:
https://github.com/astral-sh/ruff/issues/7455#issuecomment-1739802958.

## Test Plan

`cargo test`

crates/ruff_python_parser/src/lexer.rs

@@ -261,10 +261,14 @@ impl<'source> Lexer<'source> {
             'x' | 'o' | 'b'
         ));
 
+        // Lex the portion of the token after the base prefix (e.g., `9D5` in `0x9D5`).
         let mut number = LexedText::new(self.offset(), self.source);
         self.radix_run(&mut number, radix);
 
-        let value = match Int::from_str_radix(number.as_str(), radix.as_u32()) {
+        // Extract the entire number, including the base prefix (e.g., `0x9D5`).
+        let token = &self.source[self.token_range()];
+
+        let value = match Int::from_str_radix(number.as_str(), radix.as_u32(), token) {
             Ok(int) => int,
             Err(err) => {
                 return Err(LexicalError {
@@ -1462,7 +1466,8 @@ def f(arg=%timeit a = b):
 
     #[test]
     fn test_numbers() {
-        let source = "0x2f 0o12 0b1101 0 123 123_45_67_890 0.2 1e+2 2.1e3 2j 2.2j 000";
+        let source =
+            "0x2f 0o12 0b1101 0 123 123_45_67_890 0.2 1e+2 2.1e3 2j 2.2j 000 0x995DC9BBDF1939FA";
         assert_debug_snapshot!(lex_source(source));
     }
 

crates/ruff_python_parser/src/snapshots/ruff_python_parser__lexer__tests__numbers.snap

@@ -77,8 +77,14 @@ expression: lex_source(source)
         },
         60..63,
     ),
+    (
+        Int {
+            value: 0x995DC9BBDF1939FA,
+        },
+        64..82,
+    ),
     (
         Newline,
-        63..63,
+        82..82,
     ),
 ]