Describe the Bug

Description

When using import X as Y syntax, Pyrefly correctly highlights the alias (Y) but incorrectly highlights the original name (X) as an unknown/generic object instead of preserving its semantic type (class, function, etc.).

Steps to Reproduce

1. Create two Python files as shown below
2. Open test_imports.py in VSCode with Pyrefly enabled
3. Use "Developer: Inspect Editor Tokens and Scopes" on the statement from my_module import MyClass as MC
4. Inspect the token for MyClass (the original name before as)

File 1: my_module.py

class MyClass:
    """A sample class."""
    def __init__(self, value: int):
        self.value = value

def my_function() -> str:
    return "Hello"

MY_CONSTANT = 42

File 2: test_imports.py

from my_module import MyClass
from my_module import MyClass as MC
from my_module import my_function as mf

Expected Behavior

Both MyClass and MC should receive the same semantic highlighting (class type), similar to how Pylance handles this case.

In the import statement from my_module import MyClass as MC:

• MyClass should be highlighted as a class
• MC should be highlighted as a class

Actual Behavior

• MC is correctly highlighted as a class ✓
• MyClass is highlighted as an unknown/generic object ✗

When inspecting tokens:

• MC shows correct semantic token type
• MyClass shows generic/unknown semantic token type

Environment

• VSCode version: [inserisci la tua versione]
• Pyrefly version: [inserisci la tua versione]
• OS: [il tuo sistema operativo]

Comparison with Pylance

Pylance correctly highlights both the original name and the alias with their appropriate semantic types.

Additional Context

This issue affects code readability, especially in __init__.py files where the import X as X pattern is commonly used for explicit re-exports and type checking compatibility.

Sandbox Link

No response

(Only applicable for extension issues) IDE Information

VSCode version: 1.107.1
Pyrefly version: 0.46.3
OS: Windows 11