How do people handle naming for path strings versus path objects?

In modern Python, in code under my control, I convert to a pathlib.Path as soon as I can and always pass that around. This addresses the naming ambiguity you describe, gives you a richer API, and makes it immediately clear in type annotations what the value is.

def replacestringinfile(file: Path, original: str, replacement: str) -> None:
  ...

UPLOADDIR: Path = ...
@app.get("/something")
def something(path: Annotated[str, Query()]) -> None:
  if "/" in path:
    raise ValueError("invalid path")
  uploadpath = UPLOADDIR / path
  ...

I don't personally like an ambiguous str | Path parameter, in part because I often enough want to write code that can operate on a file content. So, is the "string" form "foo.csv", or "col1,col2\n1,2,3\n"? You'll also inevitably wind up immediately normalizing it to either a string or a Path immediately so you don't have duplicated code paths, and it'd be better to name the specific type you need up front.