fix: fix publish issue

* chore: test uv and ruff

* feat: new feature( universe define and download data)

* fix: fix some olg bugs

* fix: fix pytest

---------

Co-authored-by: min <[email protected]>

Author: ppmina

.coverage

@@ -96,6 +96,19 @@ convention = "google"
 
 [tool.mypy]
 python_version = "3.12"
+# 忽略缺少类型存根的第三方库
+ignore_missing_imports = true
+# 当遇到缺少类型存根的导入时，不报告错误
+warn_unused_ignores = true
+# 显示错误代码
+show_error_codes = true
+# 严格可选类型检查
+strict_optional = true
+
+# 为特定模块配置
+[[tool.mypy.overrides]]
+module = ["binance.*", "websockets.*"]
+ignore_missing_imports = true
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]
@@ -106,11 +119,5 @@ addopts = "-v --cov=src/cryptoservice --cov-report=term-missing"
 packages = ["src/cryptoservice"]
 
 [dependency-groups]
-dev = [
-    "mypy>=1.16.0",
-    "pandas-stubs>=2.2.3.250527",
-]
-test = [
-    "pytest>=8.3.5",
-    "pytest-cov>=6.1.1",
-]
+dev = ["mypy>=1.16.0", "pandas-stubs>=2.2.3.250527"]
+test = ["pytest>=8.3.5", "pytest-cov>=6.1.1"]

pyproject.toml

@@ -105,7 +105,7 @@ def store_kdtv_data(
                     values=feature,
                     index="K",
                     columns="T",
-                    aggfunc="first",
+                    aggfunc="mean",
                 )
                 array = pivot_data.values
 
@@ -129,7 +129,7 @@ def store_kdtv_data(
                     values=feature,
                     index="K",
                     columns="T",
-                    aggfunc="first",
+                    aggfunc="mean",
                 )
                 array = pivot_data.values
                 save_path = data_path / freq / feature / f"{date}.npy"
@@ -193,8 +193,8 @@ def read_kdtv_data(
             data_path = StorageUtils._resolve_path(data_path)
 
             # 生成日期范围
-            dates = pd.date_range(start=start_date, end=end_date, freq="D")
-            dates = [d.strftime("%Y%m%d") for d in dates]
+            date_range = pd.date_range(start=start_date, end=end_date, freq="D")
+            dates = [d.strftime("%Y%m%d") for d in date_range]
 
             # 读取交易对列表
             universe_path = data_path / freq / "universe_token.pkl"
@@ -230,9 +230,9 @@ def read_kdtv_data(
                         index=symbols[: len(array)],
                         columns=times,
                     )
-                    df = df.stack()  # 将时间转为索引
-                    df.name = feature
-                    date_data.append(df)
+                    stacked_series = df.stack()  # 将时间转为索引
+                    stacked_series.name = feature
+                    date_data.append(stacked_series)
 
                 if date_data:
                     # 合并同一天的所有特征
@@ -302,11 +302,14 @@ def read_and_visualize_kdtv(
                     table.add_column(col, justify="right")
 
                 # 添加数据行
-                for (symbol, time), row in df.iterrows():
-                    values = [
-                        f"{x:.4f}" if isinstance(x, (float, np.floating)) else str(x) for x in row
-                    ]
-                    table.add_row(str(time), symbol, *values)
+                for idx, row in df.iterrows():
+                    if isinstance(idx, tuple) and len(idx) == 2:
+                        symbol, time = idx
+                        values = [
+                            (f"{x:.4f}" if isinstance(x, (float, np.floating)) else str(x))
+                            for x in row
+                        ]
+                        table.add_row(str(time), str(symbol), *values)
 
                 StorageUtils.console.print(table)
             else:

src/cryptoservice/data/storage_utils.py

@@ -307,20 +307,20 @@ def from_dict(cls, data: dict[str, Any]) -> "UniverseDefinition":
             period_end_ts = snap.get("period_end_ts")
 
             if period_start_ts is None or period_end_ts is None:
-                from datetime import datetime
+                from datetime import datetime as dt_class
 
                 # 自动计算时间戳
                 period_start_ts = str(
                     int(
-                        datetime.strptime(
+                        dt_class.strptime(
                             f"{period_start_date} 00:00:00", "%Y-%m-%d %H:%M:%S"
                         ).timestamp()
                         * 1000
                     )
                 )
                 period_end_ts = str(
                     int(
-                        datetime.strptime(
+                        dt_class.strptime(
                             f"{period_end_date} 23:59:59", "%Y-%m-%d %H:%M:%S"
                         ).timestamp()
                         * 1000