ksaa 2025.9b2__py3-none-any.whl → 2025.9b3__py3-none-any.whl

kaa/db/drink.py

@@ -0,0 +1,84 @@
+from dataclasses import dataclass
+
+from .sqlite import select, select_many
+
+@dataclass
+class Drink:
+    """饮品"""
+    id: str
+    asset_id: str
+    name: str
+
+    @classmethod
+    def from_asset_id(cls, asset_id: str) -> 'Drink | None':
+        """
+        根据 asset_id 查询 Drink。
+        """
+        row = select("""
+        SELECT
+            id,
+            assetId,
+            name
+        FROM ProduceDrink
+        WHERE assetId = ?;
+        """, asset_id)
+        if row is None:
+            return None
+        id, asset_id, name = row
+        return cls(id, asset_id, name)
+    
+    @classmethod
+    def all(cls) -> list['Drink']:
+        """获取所有饮品"""
+        rows = select_many("""
+        SELECT
+            id,
+            assetId,
+            name
+        FROM ProduceDrink;
+        """)
+        results = []
+        for row in rows:
+            id, asset_id, name = row
+            results.append(cls(id, asset_id, name))
+        return results
+    
+    @classmethod
+    def ordinary_drinks_name(cls) -> list[str]:
+        """获取所有平凡的（不需要额外操作）的饮料"""
+        return [
+            '初星水', # [kaa/resources/drinks/img_general_pdrink_1-001.png]
+            '烏龍茶', # [kaa/resources/drinks/img_general_pdrink_1-004.png]
+            'ミックススムージー', # [kaa/resources/drinks/img_general_pdrink_2-001.png]
+            'リカバリドリンク', # [kaa/resources/drinks/img_general_pdrink_2-003.png]
+            'フレッシュビネガー', # [kaa/resources/drinks/img_general_pdrink_2-004.png]
+            'ブーストエキス', # [kaa/resources/drinks/img_general_pdrink_2-008.png]
+            'パワフル漢方ドリンク', # [kaa/resources/drinks/img_general_pdrink_2-009.png]
+            'センブリソーダ', # [kaa/resources/drinks/img_general_pdrink_2-010.png]
+            '初星ホエイプロテイン', # [kaa/resources/drinks/img_general_pdrink_3-001.png]
+            '初星スペシャル青汁', # [kaa/resources/drinks/img_general_pdrink_3-005.png]
+            '初星スペシャル青汁X', # [kaa/resources/drinks/img_general_pdrink_3-013.png]
+            'ビタミンドリンク', # [kaa/resources/drinks/img_general_pdrink_1-002.png]
+            'アイスコーヒー', # [kaa/resources/drinks/img_general_pdrink_1-003.png]
+            'スタミナ爆発ドリンク', # [kaa/resources/drinks/img_general_pdrink_2-005.png]
+            '厳選初星マキアート', # [kaa/resources/drinks/img_general_pdrink_3-002.png]
+            '初星ブーストエナジー', # [kaa/resources/drinks/img_general_pdrink_3-004.png]
+            # '初星黒酢', # [kaa/resources/drinks/img_general_pdrink_3-012.png]
+            'ルイボスティー', # [kaa/resources/drinks/img_general_pdrink_1-006.png]
+            'ホットコーヒー', # [kaa/resources/drinks/img_general_pdrink_1-008.png]
+            'おしゃれハーブティー', # [kaa/resources/drinks/img_general_pdrink_2-006.png]
+            '厳選初星ティー', # [kaa/resources/drinks/img_general_pdrink_3-006.png]
+            '厳選初星ブレンド', # [kaa/resources/drinks/img_general_pdrink_3-007.png]
+            '特製ハツボシエキス', # [kaa/resources/drinks/img_general_pdrink_3-010.png]
+            'ジンジャーエール', # [kaa/resources/drinks/img_general_pdrink_1-009.png]
+            'ほうじ茶', # [kaa/resources/drinks/img_general_pdrink_1-010.png]
+            # 'ほっと緑茶', # [kaa/resources/drinks/img_general_pdrink_2-007.png]
+            '厳選初星チャイ', # [kaa/resources/drinks/img_general_pdrink_3-008.png]
+            '初星スーパーソーダ', # [kaa/resources/drinks/img_general_pdrink_3-009.png]
+            '初星湯', # [kaa/resources/drinks/img_general_pdrink_3-011.png
+        ]
+
+if __name__ == '__main__':
+    from pprint import pprint as print
+    print(Drink.from_asset_id('img_general_pdrink_1-001'))
+    print([(str(drink.name) + '\', # [kaa/resources/drink/' + str(drink.asset_id) + '.png]') for drink in Drink.all()])

kaa/db/sqlite.py

@@ -1,25 +1,42 @@
-import os
 import sqlite3
-from typing import Any, cast
+import threading
+from logging import getLogger
+from typing import Any, cast, Dict, List, Optional
 
 from kaa import resources as res
 
-_db: sqlite3.Connection | None = None
 _db_path = cast(str, res.__path__)[0] + '/game.db'
 
-def select_many(query: str, *args) -> list[Any]:
-    global _db
-    if not _db:
-        _db = sqlite3.connect(_db_path)
-    c = _db.cursor()
+_db_dict = {}
+
+logger = getLogger(__name__)
+
+def _ensure_db() -> sqlite3.Connection:
+    """
+    确保数据库连接已建立
+    培育过程是新开线程，不同线程的connection不能使用
+    # TODO 培育结束需要关闭connection
+    """
+    global _db_dict
+    thread_id = threading.current_thread().ident
+    if thread_id not in _db_dict:
+        _db_dict[thread_id] = sqlite3.connect(_db_path)
+        _db_dict[thread_id].row_factory = sqlite3.Row
+        logger.info("Database connection established for thread: %s", thread_id)
+    return _db_dict[thread_id]
+
+
+def select_many(query: str, *args) -> List[Dict[str, Any]]:
+    """执行查询并返回多行结果，每行为字典格式"""
+    db = _ensure_db()
+    c = db.cursor()
     c.execute(query, args)
     return c.fetchall()
 
 
-def select(query: str, *args) -> Any:
-    global _db
-    if not _db:
-        _db = sqlite3.connect(_db_path)
-    c = _db.cursor()
+def select(query: str, *args) -> Optional[Dict[str, Any]]:
+    """执行查询并返回单行结果，为字典格式"""
+    db = _ensure_db()
+    c = db.cursor()
     c.execute(query, args)
     return c.fetchone()

kaa/game_ui/drinks_overview.py

@@ -0,0 +1,166 @@
+import logging
+from collections import deque
+
+import cv2
+import numpy as np
+from kotonebot import action
+from cv2.typing import MatLike
+from kotonebot.primitives import RectTuple
+
+from kaa.tasks import R
+from kaa.util import paths
+from kaa.db.drink import Drink
+from kaa.image_db import ImageDatabase, HistDescriptor, FileDataSource
+
+logger = logging.getLogger(__name__)
+_db: ImageDatabase | None = None
+
+def preprocess_drink_slot_img(img: MatLike) -> MatLike:
+    """预处理饮品图像，使得图像识别结果更正确
+    
+    :param img: 输入的饮品槽图像，大小 68x68 (BGR)
+    :return: 处理后的图像，大小 68x68，其中会出现更大量的纯白色，便于识别
+    """
+    PREPROCESS_COLOR_THRESHOLD = 250
+
+    assert img.shape[2] == 3
+    h, w, _ = img.shape
+
+    # 分通道
+    b, g, r = cv2.split(img)
+
+    # 把 b==255 的像素修正为纯白
+    mask = (b >= PREPROCESS_COLOR_THRESHOLD)
+    g[mask] = 255
+    b[mask] = 255
+    r[mask] = 255
+    img = cv2.merge([b, g, r])
+
+    # BFS 把边缘连通区域染成纯白
+    visited = np.zeros((h, w), dtype=bool)
+    q = deque()
+
+    # 从边缘像素开始
+    for x in range(w):
+        q.append((0, x))
+        q.append((h - 1, x))
+    for y in range(h):
+        q.append((y, 0))
+        q.append((y, w - 1))
+    
+    # 右上角禁止传播，因为一些饮料的管子会插到圈圈外面，导致白色泄露
+    right_top_x = w / 2
+    right_top_y = h / 4
+
+    while q:
+        y, x = q.popleft()
+        if not (0 <= x < w and 0 <= y < h):
+            continue
+        if visited[y, x]:
+            continue
+        visited[y, x] = True
+
+        # 设置为白色
+        img[y, x] = [255, 255, 255]
+
+        # 四邻域扩展
+        for dy, dx in [(-1,0), (1,0), (0,-1), (0,1)]:
+            ny, nx = y + dy, x + dx
+            if 0 <= nx < w and 0 <= ny < h and not (nx > right_top_x and ny < right_top_y) and not visited[ny, nx] and not np.all(img[ny, nx] >= PREPROCESS_COLOR_THRESHOLD):
+                q.append((ny, nx))
+
+    return img
+
+def drinks_db() -> ImageDatabase:
+    global _db
+    if _db is None:
+        logger.info('Loading drinks database...')
+        path = paths.resource('drinks')
+        db_path = paths.cache('drinks.pkl')
+        _db = ImageDatabase(FileDataSource(str(path)), db_path, HistDescriptor(8), name='drinks')
+    return _db
+
+def match_first_drinks(img: MatLike, delta_threshold: float = 0.7) -> Drink | None:
+    """
+    将给定图像与所有饮品 进行匹配，并返回最接近的一个饮品
+
+    :img: 输入图像，格式为 BGR 68x68 左右。
+    :delta_threshold: 距离差值阈值，如果第1个匹配结果和第2个匹配结果的 距离的差值 超过delta_threshold，那么才视为匹配成功。
+    :return: 若匹配成功，则返回 饮品的信息 和 前1~3个匹配饮品的数据库查询结果（用于更详细的日志输出）；否则返回 None。
+    """
+
+    img = preprocess_drink_slot_img(img)
+    
+    db = drinks_db()
+    matches = db.match_all(img, threshold=114514)
+    if len(matches) == 0:
+        return None
+
+    asset_id = matches[0].key.strip().split('.')[0]
+    drink = Drink.from_asset_id(asset_id)
+    if drink is None:
+        return None
+
+    if len(matches) == 1:
+        logger.info("Only 1 drink match result: %s", str(drink.name))
+        return drink
+    
+    matches = matches[:3]
+    matches_distance = [round(m.distance, 2) for m in matches]
+
+    if matches[1].distance - matches[0].distance <= delta_threshold:
+        logger.info(
+            "Mismatched drink: %s. First 3 matches distance: %s. Delta: %s",
+            str(drink.name),
+            str(matches_distance),
+            str(round(matches[1].distance - matches[0].distance, 2))
+        )
+        return None
+
+    logger.info(
+        "Matched drink: %s. First 3 matches distance: %s. Delta: %s",
+        str(drink.name),
+        str(matches_distance),
+        str(round(matches[1].distance - matches[0].distance, 2))
+    )
+    return drink
+
+@action('定位考试中出现的所有饮品', screenshot_mode='manual')
+def locate_all_drinks_in_3_drink_slots(img: MatLike) -> list[tuple[Drink, RectTuple]]:
+    """
+    将当前图像的左下角的3个饮品槽与所有饮品 进行匹配，并返回匹配上的所有饮品和他们的位置
+
+    :img: 输入图像，格式为 BGR 720x1280。
+    :return: 若匹配成功，则返回饮品的信息，否则返回 None。
+    """
+    
+    potential_rects: list[RectTuple] = [
+        R.InPurodyuusu.BoxDrink1.rect,
+        R.InPurodyuusu.BoxDrink2.rect,
+        R.InPurodyuusu.BoxDrink3.rect
+    ]
+
+    results = list[tuple[Drink, RectTuple]]()
+
+    for rect in potential_rects:
+        x, y, w, h = rect
+        img_slot = img[y:y+h, x:x+w]
+
+        drink = match_first_drinks(img_slot)
+
+        if drink is not None:
+            results.append((drink, rect))
+
+    return results
+
+if __name__ == '__main__':
+    from logging import getLogger
+    logging.basicConfig(level=logging.INFO, format='[%(asctime)s] [%(levelname)s] [%(name)s] %(message)s')
+
+    img = R.InPurodyuusu.Screenshot5Cards.data
+    # img = R.InPurodyuusu.ScreenshotSenseiTipConsult.data
+    # img = R.InPurodyuusu.Screenshot1Cards.data
+    # img = R.InPurodyuusu.ScreenshotDrinkTest.data # len = 2
+    # img = R.InPurodyuusu.Screenshot4Cards.data # len = 0
+    results = locate_all_drinks_in_3_drink_slots(img)
+    print(len(results), ":", results)

kaa/game_ui/idols_overview.py

@@ -141,11 +141,17 @@ def locate_idol(skin_id: str) -> Rect | None:
 
     sc.update()
     logger.debug('Idol preview pages count: %s', repr(sc.page_count))
-    pc = sc.page_count
-    assert pc is not None
+
+    if sc.page_count is not None:
+        iterator = sc(4 / (sc.page_count * 12) * 0.8)
+    else:
+        # 没找到ScrollBar的情况，只执行一次for循环
+        logger.warning('Not found ScrollBar in Idol overview page.')
+        iterator = range(1)
+
     # 1280x720 分辨率下，一行 4 个，一页共 12 个。
     # 一次只翻 0.8 行。
-    for _ in sc(4 / (pc * 12) * 0.8):
+    for _ in iterator:
         img = device.screenshot()
         # 只保留 BoxIdolOverviewIdols 区域
         mask = np.zeros_like(img)

kaa/image_db/db.py

@@ -95,7 +95,19 @@ class ImageDatabase:
         # 载入数据源
         logger.debug('Loading data source...')
         for key, value in self.source:
-            self.insert(key, value)
+            try:
+                self.insert(key, value)
+            except Exception as e:
+                logger.error(
+                    "\n"
+                    "Error inserting key: %s\n"
+                    "Error message: %s\n"
+                    "资源可能损坏，请检查并删除 `kaa/resources/idol_cards` 下的损坏文件，"
+                    "然后重新执行 `tools/db/extract_resources.py`",
+                    key,
+                    str(e).strip()
+                )
+                raise # 继续抛异常，让程序崩溃
         self.save()
         
     @property
@@ -134,7 +146,7 @@ class ImageDatabase:
         for name, image in images.items():
             self.insert(name, image, overwrite=overwrite)
 
-    def search(self, query: MatLike, threshold: float = 10) -> list[DatabaseQueryResult]:
+    def match_all(self, query: MatLike, threshold: float = 10) -> list[DatabaseQueryResult]:
         """
         搜索图片，返回所有符合阈值要求的图片，并按相似度降序排序。
 
@@ -149,6 +161,14 @@ class ImageDatabase:
             if dist < threshold:
                 results.append(DatabaseQueryResult(key, feature, float(dist)))
         results.sort(key=lambda x: x.distance)
+
+        # 可视化
+        # print("MinDist = ", results[0].distance, results[1].distance, results[2].distance)
+        # cv2.imshow("query", query)
+        # # cv2.imshow("query_feature", query_feature)
+        # cv2.waitKey(0)
+        # cv2.destroyAllWindows()
+
         return results
 
     def match(self, query: MatLike, threshold: float = 10) -> DatabaseQueryResult | None:
@@ -159,7 +179,7 @@ class ImageDatabase:
         :param threshold: 距离阈值。阈值越大，对相似度的要求越低。
         :return: 匹配结果。
         """
-        results = self.search(query, threshold)
+        results = self.match_all(query, threshold)
         if len(results) > 0:
             return results[0]
         else:

kaa/metadata.py

@@ -1,6 +1,31 @@
 WHATS_NEW = """
 # 更新日志
 ## kaa
+### v2025.09b3
+发布日期：2025-09-20
+
+【更新内容】
+1. 新增培育内考试开始时自动喝饮料（感谢 @YXHXianYu）
+2. 修复如果培育偶像选择页面展示的偶像过少，不出现滚动条时，会报错的问题（感谢 @YXHXianYu）
+3. 一些开发优化流程
+
+【完整日志】
+脚本：
+* [重构] 移除一些无用 import（#b502d95）
+* [修复] 修复选择角色页面滚动条不存在时导致程序终止的问题（#56aa9e1）
+* [新增] 在考试开始时，会自动消耗大部分饮品（#019eb9a）
+* [新增] 新增带图像预处理算法的饮品识别功能（#6b2fac1）
+
+界面：
+* [新增] UI 内支持检索测试版本的 kaa（#799bf27）
+
+其他：
+* [其他] 为资源提取脚本新增技能卡资源提取（#b80746c）
+* [其他] 为提取资源脚本新增并发下载与校验（#24dacb8）
+* [其他] 增加了资源损坏时的日志输出（#33c4ef9）
+* [其他] 在requirements.dev中锁定UnityPy的版本为1.21.3（#6bdf446）
+* [其他] 优化 justfile 从 Git tag 获取版本号的逻辑（#e7ee954）
+
 ### v2025.09b2
 发布日期：2025-09-14