What happened:
Calling dask.graph_manipulation.clone on the result of a dask.array.blockwise operation with a (iterable) literal argument results in a ValueError.

What you expected to happen:
The dask.graph_manipulation.clone operation to succeed.

Minimal Complete Verifiable Example:

import dask.array as da
from dask.graph_manipulation import clone

if __name__ == "__main__":

    x = da.ones(10)
    y = (1,)

    z = da.blockwise(lambda x, y: x + y[0], "r",
                     x, "r",
                     y, None)

    clone(z)

Anything else we need to know?:
The above produces:

TypeError: unhashable type: 'list'

I have traced the problem to:

In the above, the conditional if k in names causes the ValueError when k is not hashable.

Environment:

• Dask version: 2022.4.1
• Python version: 3.8.10
• Operating System: Ubuntu 20.04
• Install method (conda, pip, source): pip