Add regular and density-based tilers + Link model grid search

[piyushrpt]

• RegularTiler is meant to be used when we need to break up data into regular pieces - like snaphu
• DensityTiler is meant to be used where we need to break up data into pieces where number of good points are roughly the same - like sparse data
• Added a unit test for an EQ like signal as requested
• Added link model that uses grid search for velocity / dem error Built on top of #10

[piyushrpt]

You can test the density tiler with the following example:

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.patches import Rectangle
from spurt.utils import DensityTiler

shape = (5000, 7000)
ii = np.random.beta(1, 1, size=50000)
jj = np.random.beta(1, 1, size=50000)
jj = ii * 0.7 + jj * 0.3
ii = ii * shape[0]
jj = jj * shape[1]

points = np.column_stack((ii, jj))

tiler = DensityTiler(points, shape=shape, max_tiles=16)
print(tiler.tiles.shape, tiler.neighbors.shape)

fig, ax = plt.subplots(1, 1, figsize=(7, 7))
ax.set_ylim(0, shape[0])
ax.set_xlim(0, shape[1])
sum_diag = 0
sum_diag2 = 0
for p in tiler.tiles:
    ax.add_patch(Rectangle((p[1], p[0]), p[3] - p[1], p[2] - p[0],
                           facecolor='none', linewidth=1, edgecolor='k'))
    ax.scatter(jj, ii, color='grey',alpha=0.1)

plt.show()