Peak outside omega window error

[gustafsonsven]

Say I have an image series cache file that is loaded with the simple loader and has metadata which says omega goes from 0 to 360. I then alter the omega bounds in the loader to be 0 to 359.5 I get the expected warning:

which I click yes to.

I then run fit grains and warning such as this:

get tossed out but fitting still works (good). I just want to know why such warnings are being thrown, can we remove them/are they indicating something is wrong?

[psavery]

So, this warning is coming from here.

It shows up if you have any range of omega values that aren't associated with a frame, and a
predicted spot falls close to that missing range (within omega tolerance, specifically). In
your case, the missing range is 359.5 - 360 (aka 0) degrees. The code does take wrapping into
account, so 360.1 degrees, for example, wraps back to 0.1 degrees and uses the first frame.
But those spots for [-1 3 -1] and [1 -3 -1] must have predicted omegas close to 359.5 or 0,
within omega tolerance, and hence the warning occurred.

The predicted spot just gets skipped, so you can likely still end up with a good fit if you
have plenty of other spots.

We can potentially make this a debug() instead of a warning() so it won't show up for most
users, if you think it's safe to normally not show it.

Something else we can potentially do is only skip the invalid frames when they occur on the
edges of the window. For example, if it's looking for a spot on frames with omega values of
[359.0, 359.2, 359.4, 359.6], right now it skips the entire prediction because the last
omega, 359.6, doesn't belong to any frame. But it could just drop that one and search for
the spot using the first three frames. That could potentially help you get an even better
fit. Maybe something we should discuss tomorrow.

[psavery]

I talked with @gustafsonsven and we decided to keep the current behavior of skipping the spot, but make this a debug() message so it wouldn't normally show up for users.