增加去除每天第一个盲盒，以及增加模拟投盲盒

.gitignore

@@ -10,4 +10,6 @@ __pycache__/
 # Ignore environment files
 .env
 .venv/
-venv/
+venv/
+
+data/

blindbox_analysis.py

@@ -2,7 +2,7 @@ import argparse
 from datetime import timedelta
 import os
 import traceback
-from typing import Any, TypedDict
+from typing import Any, Optional, TypedDict
 import pandas as pd
 
 from danmaku_reader import DanmakuReader
@@ -12,9 +12,15 @@ class BlindboxItemData(TypedDict):
     price: int
     probability: float
 
+class PityItemData(TypedDict):
+    threshold: int
+    item_name: str
+    price: int
+
 class BlindboxInfo(TypedDict):
     price: int
     items: dict[str, BlindboxItemData]
+    pity: Optional[PityItemData]
 
 blindbox_config: dict[str, BlindboxInfo] = {
     "星月盲盒": {
@@ -50,6 +56,66 @@ blindbox_config: dict[str, BlindboxInfo] = {
             }
         }
     },
+    "幸运盲盒": {
+        "price": 50,
+        "items": {
+            "幸运泡泡": {
+                "price": 15,
+                "probability": 49.6
+            },
+            "星光铃铛": {
+                "price": 52,
+                "probability": 42.2
+            },
+            "梦雾纸签": {
+                "price": 100,
+                "probability": 5
+            },
+            "福灵小兽": {
+                "price": 200,
+                "probability": 2.4
+            },
+            "星愿花园": {
+                "price": 600,
+                "probability": 0.8
+            }
+        },
+        "pity": {
+            "threshold": 66,
+            "item_name": "糖梦福狸",
+            "price": 200
+        }
+    },
+    "幸运盲盒S": {
+        "price": 500,
+        "items": {
+            "初兆光符": {
+                "price": 160,
+                "probability": 54.04
+            },
+            "福引转轮": {
+                "price": 520,
+                "probability": 40
+            },
+            "光羽预言": {
+                "price": 1000,
+                "probability": 4.8
+            },
+            "幽镜之门": {
+                "price": 5000,
+                "probability": 1
+            },
+            "命契幻境": {
+                "price": 30000,
+                "probability": 0.16
+            }
+        },
+        "pity": {
+            "threshold": 100,
+            "item_name": "幻梦晶球",
+            "price": 3000
+        }
+    },
     "心动盲盒": {
         "price": 150,  # 请根据实际盲盒价格填写
         "items": {
@@ -250,13 +316,33 @@ def profit_statistic(df: pd.DataFrame) -> dict[str, dict[str, Any]]:
 def run_statistics(df: pd.DataFrame) -> dict[str, dict[str, Any]]:
     """运行统计分析"""
     stats = {}
+
     # 统计整体数据
     total_stats = profit_statistic(df)
     stats["整体"] = total_stats
+
+    # 统计去掉每个用户每天第一次开盲盒的数据
+    def remove_first_record_per_group(group):
+        """去掉每组的第一条记录"""
+        return group.iloc[1:]  # 从第二条开始返回
+
+    # 确保时间列是datetime格式
+    if 'time' in df.columns:
+        df['time'] = pd.to_datetime(df['time'])
+        # 提取日期部分（去掉时间）
+        df['date'] = df['time'].dt.date
+    # 按日期、用户ID、盲盒名称分组
+    grouped = df.groupby(['date', 'username', 'blindbox_name'])
+    # 去掉每组的第一条数据
+    df_filtered = grouped.apply(remove_first_record_per_group).reset_index(drop=True)
+    filtered_stats = profit_statistic(df_filtered)
+    stats["去除保底"] = filtered_stats
+
     # 统计周五以外的数据
     non_friday_df = df[pd.to_datetime(df["time"]).dt.weekday != 4]
     non_friday_stats = profit_statistic(non_friday_df)
     stats["非周五"] = non_friday_stats
+    
     # 统计周五的数据
     friday_df = df[pd.to_datetime(df["time"]).dt.weekday == 4]
     friday_stats = profit_statistic(friday_df)
@@ -336,4 +422,6 @@ if __name__ == "__main__":
         # 运行统计分析
         stats = run_statistics(result_df)
         print("\n统计结果:")
-        print_tree(stats)
+        print_tree(stats)
+    else:
+        print("未找到任何盲盒数据")

blindbox_simulation.py

@@ -0,0 +1,69 @@
+import numpy as np
+import pandas as pd
+from blindbox_analysis import blindbox_config, print_tree, profit_statistic
+
+def generate_random_data(blindbox_name: str, num_records: int):
+    """
+    生成随机盲盒数据
+    """
+    np.random.seed(42)  # 固定随机种子，确保结果可复现
+
+    dataset = []
+    for i in range(num_records):
+        # 随机生成用户ID和用户名
+        uid = np.random.randint(1000, 2000)
+        username = f'user_{uid}'
+
+        blindbox_info = blindbox_config[blindbox_name]
+        pity_info = blindbox_info.get('pity', None)
+
+        # 抽选盲盒物品
+        rand_num = np.random.rand()
+        sum_prob = 0
+        selected_item = None
+        for item, item_info in blindbox_info['items'].items():
+            prob = item_info['probability'] / 100
+            if rand_num > sum_prob and rand_num <= sum_prob + prob:
+                selected_item = (item, item_info)
+                break
+            sum_prob += prob
+
+        if selected_item is None:
+            raise ValueError("未能选中任何物品")
+
+        # 构建随机数据记录
+        dataset.append({
+            'uid': uid,
+            'username': username,
+            'blindbox_name': blindbox_name,
+            'blindbox_price': blindbox_info['price'],
+            'item_name': selected_item[0],
+            'item_price': selected_item[1]['price'],
+            'profit': selected_item[1]['price'] - blindbox_info['price'],
+            'time': pd.Timestamp.now()
+        })
+
+        if pity_info:
+            # 模拟保底机制
+            if (i + 1) % pity_info['threshold'] == 0:
+                dataset.append({
+                    'uid': uid,
+                    'username': username,
+                    'blindbox_name': blindbox_name,
+                    'blindbox_price': blindbox_info['price'],
+                    'item_name': pity_info['item_name'],
+                    'item_price': pity_info['price'],
+                    'profit': pity_info['price'] - blindbox_info['price'],
+                    'time': pd.Timestamp.now()
+                })
+
+    return pd.DataFrame(dataset)
+
+if __name__ == "__main__":
+    # 生成随机数据
+    stats = {}
+    for box_name in blindbox_config.keys():
+        df = generate_random_data(box_name, 10000)
+        stats[box_name] = profit_statistic(df)
+
+    print_tree(stats)

requirements.txt

@@ -1 +1,2 @@
-pandas
+pandas
+numpy

requirements[asr].txt

@@ -0,0 +1,4 @@
+torch 
+torchaudio
+--index-url https://download.pytorch.org/whl/cu126
+funasr