Move all geometries impls into impl Distance for Euclidean

As part of aligning all our line measure traits (see
#1181), all the Point x Point
distance calculations had been moved into `line_measures::metric_spaces`.

Euclidean space is the only space that implements distance calculations
on *non* point geometries. This commit moves all those calculations onto
the `line_measures::metric_spaces::Euclidean` so that we can fully
deprecate the existing EuclideanDistance trait.

geo/benches/euclidean_distance.rs

@@ -1,5 +1,5 @@
 use criterion::{criterion_group, criterion_main};
-use geo::algorithm::{ConvexHull, EuclideanDistance};
+use geo::algorithm::{ConvexHull, Distance, Euclidean};
 use geo::{polygon, Polygon};
 
 fn criterion_benchmark(c: &mut criterion::Criterion) {
@@ -37,9 +37,7 @@ fn criterion_benchmark(c: &mut criterion::Criterion) {
             (x: -6.064453, y: 68.49604),
         ];
         bencher.iter(|| {
-            criterion::black_box(
-                criterion::black_box(&poly1).euclidean_distance(criterion::black_box(&poly2)),
-            );
+            criterion::black_box(Euclidean::distance(&poly1, &poly2));
         });
     });
 
@@ -81,9 +79,7 @@ fn criterion_benchmark(c: &mut criterion::Criterion) {
             ]
             .convex_hull();
             bencher.iter(|| {
-                criterion::black_box(
-                    criterion::black_box(&poly1).euclidean_distance(criterion::black_box(&poly2)),
-                );
+                criterion::black_box(Euclidean::distance(&poly1, &poly2));
             });
         },
     );

geo/src/algorithm/concave_hull.rs

@@ -1,8 +1,8 @@
 use crate::convex_hull::qhull;
 use crate::utils::partial_min;
 use crate::{
-    coord, Centroid, Coord, CoordNum, Euclidean, EuclideanDistance, GeoFloat, Length, Line,
-    LineString, MultiLineString, MultiPoint, MultiPolygon, Point, Polygon,
+    coord, Centroid, Coord, CoordNum, Distance, Euclidean, GeoFloat, Length, Line, LineString,
+    MultiLineString, MultiPoint, MultiPolygon, Point, Polygon,
 };
 use rstar::{RTree, RTreeNum};
 use std::collections::VecDeque;
@@ -134,8 +134,8 @@ where
             let closest_point =
                 candidates.fold(Point::new(point.x, point.y), |acc_point, candidate| {
                     let candidate_point = Point::new(candidate.x, candidate.y);
-                    if line.euclidean_distance(&acc_point)
-                        > line.euclidean_distance(&candidate_point)
+                    if Euclidean::distance(&line, &acc_point)
+                        > Euclidean::distance(&line, &candidate_point)
                     {
                         candidate_point
                     } else {
@@ -154,8 +154,8 @@ where
             let closest_edge_option = match peeked_edge {
                 None => None,
                 Some(&edge) => Some(edges_nearby_point.fold(*edge, |acc, candidate| {
-                    if closest_point.euclidean_distance(&acc)
-                        > closest_point.euclidean_distance(candidate)
+                    if Euclidean::distance(&closest_point, &acc)
+                        > Euclidean::distance(&closest_point, candidate)
                     {
                         *candidate
                     } else {
@@ -164,8 +164,8 @@ where
                 })),
             };
             let decision_distance = partial_min(
-                closest_point.euclidean_distance(&line.start_point()),
-                closest_point.euclidean_distance(&line.end_point()),
+                Euclidean::distance(&closest_point, &line.start_point()),
+                Euclidean::distance(&closest_point, &line.end_point()),
             );
             if let Some(closest_edge) = closest_edge_option {
                 let far_enough = edge_length / decision_distance > concavity;