atsa-python: need to catch bad slope calculations

[rbavery]

Some images can't get usable histograms built because of clouds (see the product of the histogram calculation, red or blue means)

[0.0, 1167.5714285714287, 1230.55, 1303.2, 1360.6, 1430.6, 1502.55, 1566.1, 1673.7, 1721.1, 1808.45, 1868.65, 4259.25]
[0.0, 1574.0, 1619.65, 1690.15, 1713.75, 1749.8, 1773.7, 1825.2, 1872.4, 1938.5, 4041.95]
[0.0, 1299.5882352941176, 1356.65, 1429.4, 1494.15, 1562.7, 1673.95, 1728.35, 1806.95, 1852.25, 1921.4, 1942.5, 1892.9, 2496.2]
[0.0, 5569.4]

So we need to average the clear sky lines for parts of the time series that are good (slope near 1.5) in order to get better HOT values for these images. See idl version for other edge cases to catch