Update dark model to correspond to SOTA 2017 analysis

[taldcroft]

DO NOT MERGE yet.

This updates the dark_model module, which has a downstream effect on acquisition probabilities via the warm fraction (get_warm_fracs()). The changes are slightly larger than initially anticipated, as much as 0.5 mag in the CCD warm limit temperature. Fortunately the changes are in a good direction (warmer allowed temps) but a bit more investigation would be warranted before promoting to flight.

However this is fine for SOTA analysis immediately (e.g. open the branch in a git repo somewhere and set PYTHONPATH).

[taldcroft]

Changes to the model warm fraction are visible in fb58a01.

[taldcroft]

Previous analysis had a mistake in forgetting to remove the zodiacal brightness. Here is a comparison of current flight dark model vs. the new dark model here. The new one is significantly more accurate. This is not surprising given the obvious slope change in the amplitude fit around 2012.

Flight dark model

2015:090  lim=  50  model:335670  data:315910  frac_err:   6.3%
2017:092  lim=  50  model:453847  data:415932  frac_err:   9.1%
2017:185  lim=  50  model:395578  data:354761  frac_err:  11.5%
2015:090  lim= 100  model:196830  data:191791  frac_err:   2.6%
2017:092  lim= 100  model:295717  data:281794  frac_err:   4.9%
2017:185  lim= 100  model:237503  data:216311  frac_err:   9.8%
2015:090  lim= 200  model: 72143  data: 74272  frac_err:  -2.9%
2017:092  lim= 200  model:143926  data:140980  frac_err:   2.1%
2017:185  lim= 200  model: 92757  data: 83158  frac_err:  11.5%
2015:090  lim=1000  model:   636  data:   577  frac_err:  10.2%
2017:092  lim=1000  model:  1995  data:  1671  frac_err:  19.4%
2017:185  lim=1000  model:   885  data:   641  frac_err:  38.1%
2013:011  t_ccd=-16.4  lim= 100  model:121130  data:119110  frac_err:   1.7%
2013:096  t_ccd=-16.2  lim= 100  model:126827  data:126420  frac_err:   0.3%
2013:191  t_ccd=-15.3  lim= 100  model:143261  data:143587  frac_err:  -0.2%
2013:281  t_ccd=-15.9  lim= 100  model:138194  data:137892  frac_err:   0.2%
2014:041  t_ccd=-14.6  lim= 100  model:163440  data:163375  frac_err:   0.0%
2014:104  t_ccd=-15.8  lim= 100  model:146655  data:142199  frac_err:   3.1%
2014:220  t_ccd=-14.8  lim= 100  model:168198  data:173490  frac_err:  -3.1%
2014:289  t_ccd=-13.9  lim= 100  model:186089  data:186412  frac_err:  -0.2%
2015:014  t_ccd=-14.9  lim= 100  model:172840  data:165513  frac_err:   4.4%
2015:090  t_ccd=-13.7  lim= 100  model:196830  data:191791  frac_err:   2.6%
2015:183  t_ccd=-14.8  lim= 100  model:180926  data:169482  frac_err:   6.8%
2015:273  t_ccd=-13.9  lim= 100  model:200600  data:185514  frac_err:   8.1%
2015:355  t_ccd=-15.1  lim= 100  model:184104  data:163632  frac_err:  12.5%
2016:048  t_ccd=-14.6  lim= 100  model:195144  data:177047  frac_err:  10.2%
2016:125  t_ccd=-13.7  lim= 100  model:215548  data:195037  frac_err:  10.5%
2016:188  t_ccd=-13.8  lim= 100  model:216125  data:197372  frac_err:   9.5%
2016:278  t_ccd=-13.0  lim= 100  model:235035  data:217033  frac_err:   8.3%
2017:010  t_ccd=-13.5  lim= 100  model:229252  data:211866  frac_err:   8.2%
2017:092  t_ccd=-10.4  lim= 100  model:295717  data:281794  frac_err:   4.9%
2017:185  t_ccd=-13.5  lim= 100  model:237503  data:216311  frac_err:   9.8%

New dark model (this PR)

2015:090  lim=  50  model:310177  data:315910  frac_err:  -1.8%
2017:092  lim=  50  model:400108  data:415932  frac_err:  -3.8%
2017:185  lim=  50  model:350236  data:354761  frac_err:  -1.3%
2015:090  lim= 100  model:185177  data:191791  frac_err:  -3.4%
2017:092  lim= 100  model:265989  data:281794  frac_err:  -5.6%
2017:185  lim= 100  model:213491  data:216311  frac_err:  -1.3%
2015:090  lim= 200  model: 68983  data: 74272  frac_err:  -7.1%
2017:092  lim= 200  model:129527  data:140980  frac_err:  -8.1%
2017:185  lim= 200  model: 81587  data: 83158  frac_err:  -1.9%
2015:090  lim=1000  model:   557  data:   577  frac_err:  -3.5%
2017:092  lim=1000  model:  1488  data:  1671  frac_err: -11.0%
2017:185  lim=1000  model:   652  data:   641  frac_err:   1.7%
2013:011  t_ccd=-16.4  lim= 100  model:124601  data:119110  frac_err:   4.6%
2013:096  t_ccd=-16.2  lim= 100  model:129091  data:126420  frac_err:   2.1%
2013:191  t_ccd=-15.3  lim= 100  model:143674  data:143587  frac_err:   0.1%
2013:281  t_ccd=-15.9  lim= 100  model:137712  data:137892  frac_err:  -0.1%
2014:041  t_ccd=-14.6  lim= 100  model:159847  data:163375  frac_err:  -2.2%
2014:104  t_ccd=-15.8  lim= 100  model:143358  data:142199  frac_err:   0.8%
2014:220  t_ccd=-14.8  lim= 100  model:161916  data:173490  frac_err:  -6.7%
2014:289  t_ccd=-13.9  lim= 100  model:177842  data:186412  frac_err:  -4.6%
2015:014  t_ccd=-14.9  lim= 100  model:164080  data:165513  frac_err:  -0.9%
2015:090  t_ccd=-13.7  lim= 100  model:185177  data:191791  frac_err:  -3.4%
2015:183  t_ccd=-14.8  lim= 100  model:169715  data:169482  frac_err:   0.1%
2015:273  t_ccd=-13.9  lim= 100  model:187110  data:185514  frac_err:   0.9%
2015:355  t_ccd=-15.1  lim= 100  model:170932  data:163632  frac_err:   4.5%
2016:048  t_ccd=-14.6  lim= 100  model:180473  data:177047  frac_err:   1.9%
2016:125  t_ccd=-13.7  lim= 100  model:198148  data:195037  frac_err:   1.6%
2016:188  t_ccd=-13.8  lim= 100  model:198015  data:197372  frac_err:   0.3%
2016:278  t_ccd=-13.0  lim= 100  model:214205  data:217033  frac_err:  -1.3%
2017:010  t_ccd=-13.5  lim= 100  model:207926  data:211866  frac_err:  -1.9%
2017:092  t_ccd=-10.4  lim= 100  model:265989  data:281794  frac_err:  -5.6%
2017:185  t_ccd=-13.5  lim= 100  model:213491  data:216311  frac_err:  -1.3%

[taldcroft]

Superceded by #37.