Created new pixel flats and bad pixel masks

Created new pixel flats and bad pixel masks post 2016 summer shut down. This summer shut down, we swapped the B2 camera with the spare camera. So, the pixel flats and bad pixel masks are going to be different for B2. For the rest three cameras, there shouldn't be much changes.

While I was making the pixel flats previously, I was making a small bug. Since I mostly work with SOS mountain reduction of spectra, I am used to set specflat to version v1_20 (aporeduce will only work with that version of specflat). So, I had changed the module files to load v1_20 as my default specflat. I was making the pixel masks using this version of specflat. Upon comparison with previous pixel flats, I saw the same dusty features still on the new camera. I became suspicious on this as B2 is an altogether new camera and it should have the same dusty features as of the previous camera. Finally caught the mistake and re-made all the pixel flats. So, from now on, remember to turn off specflat/v1_20 whenever dealing with specflat products.

For bad pixel masks, the decision was made to only update the B2 camera as that is the only new one. I browsed through a lot of bias, dark and lossy fiber flat frames to look for any obvious bad columns in the data. For three bias frames, there was couple of bad columns in the second quadrant. The lossy fiber has this feature for all the frames, although the column length is less. So, decided to flag this two columns as bad. Then I flagged all pixels which are deviant more than 10 sigma in the median dark and bias frames.

Here is the difference image of B2 pixel flats. I made a difference between old and new B2 pixel flats. All the dust features go away in the new flat.