Peeling

[nhurleywalker]

Bright sources just outside the field-of-view, or in the sidelobes, can cause the calibration stage to fail. I've introduced some work toward the logic needed to solve this (https://github.com/nhurleywalker/GLEAM-X-pipeline/blob/master/db/check_sources_vs_obsids.py, https://github.com/nhurleywalker/GLEAM-X-pipeline/blob/master/db/check_src_fov.py), but it's not yet a full implementation. The logical sequence I've been thinking of using is:

• Are we doing peeling? Y/N
  • If Y: Make a list of sources and apparent flux densities, and count the number that are above the peeling threshold (user-set, ~50 Jy) Are there any sources that lie above the threshold?
    • If Y: How many of those sources lie inside the field-of-view?
      • If all of them: don't worry, the calibration will have taken care of it
      • If one lies outside: Does that source have a greater apparent flux density than the source(s) inside the field-of-view?
        • If Y: peel the source that lies outside the field-of-view.
        • If N: Can't peel because it won't converge or will subtract the wrong source: could try making the image larger (up to some threshold). Does the source now lie inside the field-of-view?
          • If Y: image with a larger image size
          • If N: Directly subtract the source rather than peeling it.

We will almost certainly have to transfer a calibration solution to a field that needs peeling, do all the peeling or subtracting, and then re-calibrate without the peeled/subtracted sources in the sky model. It's this connective work that would really benefit from proper database use -- for GLEAM I just did it manually with notes in my notebook, but that's not ideal!

Tim, assigning this to you as a long-term goal.