Fixes issues found with linear fitting guess function

[jackbdoughty]

Found issues with linear fitting whereby the guess function would not work as expected e.g not guessing a straight line and division by zero error. This fixes by instead using the already implemented polynomial guess function and passing 1 as the polynomial degree. Should test by creating an example plan and two blocks to scan against eachother from- use a python script to repeatedly write to one of the blocks to simulate data from scan. See below. Check that initial guess produced in plot window fits the data provided well.

See fitting docs for more info.

Run this separately

def linear(x, m, c):
    return m * x + c

g.set_pv("CS:SB:bob", 1, is_local=True)

while True:

    val = g.get_pv("CS:SB:bob", is_local=True)
    g.set_pv("CS:SB:alice", linear(val, 0, 1),is_local=True)

Use a plan that looks like this

def plan():

    alice = block_r(float, "alice")
    bob = block_rw(float, "bob", write_config=BlockWriteConfig(settle_time_s=0.5))

    lf = LiveFit(Linear.fit(), "alice", "bob")

    fig, ax = plt.subplots()
    
    @subs_decorator(
        [
            LiveTable([bob.name, alice.name]),
            LiveFitPlot(lf, ax=ax, color="r", num_points=1000),
            LivePlot(y=alice.name, x=bob.name, ax=ax, marker="x", linestyle="none"),
        ]
    )
    def _inner() -> Generator[Msg, None, None]:
        yield from ensure_connected(bob, alice)
        yield from bp.scan([alice], bob, -5, 5, num=15)
        logger.blueskylogger.info("Some useful message")
    yield from _inner()