lightcurve figure y-range too large

[mcoughlin]

Describe the bug
For some sources (stellar?), there can be upperlimits with values of >1000 or <1000 mag. These set the y-range of the very large values, and make the figure useless.

To Reproduce
Steps to reproduce the behavior:

1. Go to https://fritz.science/alerts/ztf/ZTF17aabupib
2. Scroll down to the 'Photometry and cutouts'; the figure with the ZTF lightcurve shows a y-axis spanning from 20 to -1000

Expected behavior
The y-range of the lightcurve figure should ignore any limits (or points) with mag<0 and mag>25 (?)

Screenshots
If applicable, add screenshots to help explain your problem.

Copied from fritz-marshal/fritz-beta-feedback#83