stub frechet distance fn, new section for chi2 test for spatial corre…

…lations in examples notebook

.gitignore

@@ -2,4 +2,5 @@
 .pytest_cache
 venv
 *.pyc
-**/*_field.csv
+**/*_field.csv
+notebooks/grid_chi2.ipynb

cactice/distance.py

@@ -1,4 +1,4 @@
-from typing import List
+from typing import List, Tuple
 
 
 def hamming_distance(a: List[int], b: List[int]) -> int:
@@ -13,4 +13,10 @@ def hamming_distance(a: List[int], b: List[int]) -> int:
     a_str = ''.join([str(i) for i in a])
     b_str = ''.join([str(i) for i in b])
 
-    return sum(ca != cb for ca, cb in zip(a_str, b_str))
+    return sum(ca != cb for ca, cb in zip(a_str, b_str))
+
+
+def frechet_distance(a: List[Tuple[float, float]], b: List[Tuple[float, float]]) -> float:
+    # TODO
+    # reference impl: https://gist.github.com/MaxBareiss/ba2f9441d9455b56fbc9
+    pass

notebooks/explore.ipynb

@@ -1019,6 +1019,31 @@
     }
    }
   },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "### Testing for spatial correlation\n",
+    "\n",
+    "We can use a chi-squared test to check for evidence of spatial correlation between grid cell neighborhoods at any given length scale. We can first compute neighborhoods for each cell. Next we can retrieve for each cell the equiradial band of cells displaced from it by distance `d`, then compute the Hamming distance between the central cell's neighborhood and the neighborhood of each band cell. We compute this exhaustively for all possible values of `d` for a given grid size."
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "# TODO"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   }
+  },
   {
    "cell_type": "markdown",
    "source": [