feat(single-mode): recognize training type

auto_derby/mathtools.py

@@ -4,6 +4,9 @@
 
 from typing import Tuple
 
+import cast_unknown as cast
+import numpy as np
+
 
 def linear_interpolate(a: float, b: float, pos: float) -> float:
     return a + (b - a) * pos
@@ -57,3 +60,12 @@ def vector4(
     ) -> Tuple[int, int, int, int]:
         l, t, r, b = (self.vector(i, from_) for i in rect)
         return l, t, r, b
+
+
+
+
+def distance(a: Tuple[int, ...], b: Tuple[int, ...]) -> float:
+    assert len(a) == len(b), f"length must be same, got len(a)={len(a)} len(b)={len(b)}"
+    return cast.instance(
+        np.sqrt(np.sum((np.array(a) - np.array(b)) ** 2, axis=0)), float
+    )

auto_derby/mathtools_test.py

@@ -6,3 +6,11 @@ def test_resize_proxy_issue71():
 
     res = rp.vector(-50, 466)
     assert res == -116, res
+
+
+def test_distance():
+    assert mathtools.distance((0,), (1,)) == 1
+    assert mathtools.distance((0,), (-1,)) == 1
+    assert mathtools.distance((0, 0), (1, 1)) == 1.4142135623730951
+    assert mathtools.distance((0, 0), (-1, -1)) == 1.4142135623730951
+    assert mathtools.distance((0, 0, 0), (1, 1, 1)) == 1.7320508075688772

auto_derby/single_mode/training.py

@@ -3,7 +3,7 @@
 from __future__ import annotations
 
 import os
-from typing import Tuple, Type
+from typing import Dict, Tuple, Type
 
 import cv2
 import numpy as np
@@ -18,6 +18,7 @@
 
 class g:
     training_class: Type[Training]
+    target_levels: Dict[int, int] = {}
 
 
 def _gradient(colors: Tuple[Tuple[Tuple[int, int, int], int], ...]) -> np.ndarray:
@@ -125,12 +126,27 @@ def _recognize_level(rgb_color: Tuple[int, ...]) -> int:
 
 
 class Training:
+    TYPE_SPEED: int = 1
+    TYPE_STAMINA: int = 2
+    TYPE_POWER: int = 3
+    TYPE_GUTS: int = 4
+    TYPE_WISDOM: int = 5
+
+    ALL_TYPES = (
+        TYPE_SPEED,
+        TYPE_STAMINA,
+        TYPE_POWER,
+        TYPE_GUTS,
+        TYPE_WISDOM,
+    )
+
     @staticmethod
     def new() -> Training:
         return g.training_class()
 
     def __init__(self):
         self.level = 0
+        self.type = 0
 
         self.speed: int = 0
         self.stamina: int = 0
@@ -144,6 +160,8 @@ def __init__(self):
 
     @classmethod
     def from_training_scene(cls, img: Image) -> Training:
+        rp = mathtools.ResizeProxy(img.width)
+
         self = cls.new()
         self.confirm_position = next(
             template.match(
@@ -153,8 +171,24 @@ def from_training_scene(cls, img: Image) -> Training:
                 ),
             )
         )[1]
-
-        rp = mathtools.ResizeProxy(img.width)
+        radius = rp.vector(30, 540)
+        for t, center in zip(
+            Training.ALL_TYPES,
+            (
+                rp.vector2((78, 850), 540),
+                rp.vector2((171, 850), 540),
+                rp.vector2((268, 850), 540),
+                rp.vector2((367, 850), 540),
+                rp.vector2((461, 850), 540),
+            ),
+        ):
+            if mathtools.distance(self.confirm_position, center) < radius:
+                self.type = t
+                break
+        else:
+            raise ValueError(
+                "unknown type for confirm position: %s" % self.confirm_position
+            )
 
         self.level = _recognize_level(
             tuple(cast.list_(img.getpixel(rp.vector2((10, 200), 540)), int))
@@ -264,7 +298,25 @@ def score(self, ctx: Context) -> float:
                 (7000, 1.0),
             ),
         )
-        return (spd + sta + pow + per + int_ + skill) * success_rate
+
+        target_level = g.target_levels.get(self.type, self.level)
+        target_level_score = 0
+        if self.level < target_level:
+            target_level_score += mathtools.interpolate(
+                ctx.turn_count(),
+                (
+                    (0, 5),
+                    (24, 3),
+                    (48, 2),
+                    (72, 0),
+                ),
+            )
+        elif self.level > target_level:
+            target_level_score -= (self.level - target_level) * 5
+
+        return (
+            spd + sta + pow + per + int_ + skill + target_level_score
+        ) * success_rate
 
 
 g.training_class = Training

auto_derby/single_mode/training_test.py

@@ -16,6 +16,7 @@ def test_update_by_training_scene():
     )
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_SPEED
     assert training.level == 5
     assert training.speed == 26
     assert training.stamina == 0
@@ -33,6 +34,7 @@ def test_update_by_training_scene_2():
     )
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_WISDOM
     assert training.level == 3
     assert training.speed == 6
     assert training.stamina == 0
@@ -50,6 +52,7 @@ def test_update_by_training_scene_3():
     )
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_GUTS
     assert training.level == 5
     assert training.speed == 6
     assert training.stamina == 0
@@ -67,6 +70,7 @@ def test_update_by_training_scene_4():
     )
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_GUTS
     assert training.level == 5
     assert training.speed == 7
     assert training.stamina == 0
@@ -80,6 +84,7 @@ def test_update_by_training_scene_5():
     with _test.screenshot("training_scene_5.png") as img:
 
         training = Training.from_training_scene(img)
+        assert training.type == training.TYPE_WISDOM
         assert training.level == 2
         assert training.speed == 2
         assert training.stamina == 0
@@ -97,6 +102,7 @@ def test_update_by_training_scene_issue9():
     )
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_SPEED
     assert training.level == 1
     assert training.speed == 12
     assert training.stamina == 0
@@ -114,6 +120,7 @@ def test_update_by_training_scene_issue24():
     )
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_STAMINA
     assert training.level == 1
     assert training.speed == 0
     assert training.stamina == 9
@@ -131,6 +138,7 @@ def test_update_by_training_scene_issue51():
     )
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_SPEED
     assert training.level == 5
     assert training.speed == 21
     assert training.stamina == 0
@@ -144,6 +152,7 @@ def test_update_by_training_scene_issue55():
     img = PIL.Image.open(_TEST_DATA_PATH / "training_scene_issue55.png").convert("RGB")
 
     training = Training.from_training_scene(img)
+    assert training.type == training.TYPE_SPEED
     assert training.level == 5
     assert training.speed == 30
     assert training.stamina == 0