Smooth bp

[jmoldon]

New version with Initial bandpass completed.

• I have rewritten all the previous calibration used/produced by solve_delays and do_initial_bandpass.
• Now a single dictionary keeps all the information following the table structure described in #39.
• Right now it assumes the pipeline has been run up to the requested step (so the dictionary contains what the previous steps produced and the tables are available in the calib directory). We can add additional checks to verify that they are there, and if not give an error.
• I have rewritten new smaller and general functions: run_gaincal, run_bandpass that can be called later with other parameters.
• I have included a smooth_caltable function. Now it is executed only on the initial_bandpass "ap" calibration step. I keep a backup table before the smoothing.
• Now I produce plots for all calibration tables. When gaintype is 'G' I always produce amplitude and phase only (also even if we don't solve for amplitudes), to have all the information of what is being applied.

[jradcliffe5]

Will try to check this sometime over the weekend. Bit tied up atm!

[jradcliffe5]

Sorry its taken so long Javier.

I've been having is that the tables have not been calibrating and the pipeline not working at all on a mac at the moment. I don't have much time at all at the moment to check (as in Zambia), but keep puching to this branch and I will try to get down to sorting out the problems i have this weekend I hope. I think Im just not setting something correctly.