Point graphic and line graphic should produce reasonable masks

[cmeyer]

Right now they produce empty masks. A point graphic could produce a mask of the pixel on which it sits. A line graphic could produce a rough mask of the line - although the line profile would be preferred for analytical work.

[cmeyer]

Implementation notes:

• you will be implementing the method get_mask for the point graphic and the line graphic
• here is the get_mask method for a rectangle
• masks are currently binary, meaning they are either 1 or 0.
• the get_mask function returns a numpy array with the same size as the input image
• the code should include a separate test for both line and point
• for now, the point should just cover a single pixel; the line should be 1-pixel and not have gaps (diagonally-connected pixels are ok)
• the line algorithm should be implemented with numpy; it must run fast; it should not have explicit loops
• future possibilities are that the point will have a radius and that the line will have a width (and circular or square caps)