openseries 1.5.6__tar.gz → 1.5.7__tar.gz

{openseries-1.5.6 → openseries-1.5.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: openseries
-Version: 1.5.6
+Version: 1.5.7
 Summary: Tools for analyzing financial timeseries.
 Home-page: https://github.com/CaptorAB/openseries
 License: BSD-3-Clause
@@ -20,8 +20,8 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: Office/Business :: Financial :: Investment
-Requires-Dist: holidays (>=0.30,<0.50)
-Requires-Dist: numpy (>=1.23.2,<=2.0.0)
+Requires-Dist: holidays (>=0.30,<1.0)
+Requires-Dist: numpy (>=1.23.2,<=3.0.0)
 Requires-Dist: openpyxl (>=3.1.2,<4.0.0)
 Requires-Dist: pandas (>=2.1.2,<3.0.0)
 Requires-Dist: plotly (>=5.18.0,<6.0.0)

{openseries-1.5.6 → openseries-1.5.7}/openseries/_risk.py

@@ -6,19 +6,11 @@ from math import ceil
 from typing import Union, cast
 
 from numpy import (
-    divide,
-    float64,
-    isinf,
     mean,
-    nan,
     nan_to_num,
     quantile,
     sort,
-    sqrt,
-    square,
-    std,
 )
-from numpy.typing import NDArray
 from pandas import DataFrame, Series
 
 from openseries.types import LiteralQuantileInterp
@@ -87,56 +79,3 @@ def _var_down_calc(
         clean = nan_to_num(data)
     ret = clean[1:] / clean[:-1] - 1
     return cast(float, quantile(ret, 1 - level, method=interpolation))
-
-
-def _ewma_calc(
-    reeturn: float,
-    prev_ewma: float,
-    time_factor: float,
-    lmbda: float = 0.94,
-) -> float:
-    """
-    Calculate Exponentially Weighted Moving Average volatility.
-
-    Parameters
-    ----------
-    reeturn : float
-        Return value
-    prev_ewma : float
-        Previous EWMA volatility value
-    time_factor : float
-        Scaling factor to annualize
-    lmbda: float, default: 0.94
-        Scaling factor to determine weighting.
-
-    Returns
-    -------
-    float
-        EWMA volatility value
-
-    """
-    return cast(
-        float,
-        sqrt(square(reeturn) * time_factor * (1 - lmbda) + square(prev_ewma) * lmbda),
-    )
-
-
-def _calc_inv_vol_weights(returns: DataFrame) -> NDArray[float64]:
-    """
-    Calculate weights proportional to inverse volatility.
-
-    Parameters
-    ----------
-    returns: pandas.DataFrame
-        returns data
-
-    Returns
-    -------
-    NDArray[float64]
-        Calculated weights
-
-    """
-    vol = divide(1.0, std(returns, axis=0, ddof=1))
-    vol[isinf(vol)] = nan
-    volsum = vol.sum()
-    return cast(NDArray[float64], divide(vol, volsum))

{openseries-1.5.6 → openseries-1.5.7}/openseries/datefixer.py

@@ -62,7 +62,7 @@ def holiday_calendar(
     endyear += 1
     if startyear == endyear:
         endyear += 1
-    years = list(range(int(startyear), int(endyear)))
+    years = list(range(startyear, endyear))
 
     if isinstance(countries, str) and countries in list_supported_countries():
         staging = country_holidays(country=countries, years=years)
@@ -116,15 +116,11 @@ def date_fix(
     if isinstance(fixerdate, datetime64):
         return (
             dt.datetime.strptime(str(fixerdate)[:10], "%Y-%m-%d")
-            .replace(tzinfo=dt.timezone.utc)
+            .astimezone()
             .date()
         )
     if isinstance(fixerdate, str):
-        return (
-            dt.datetime.strptime(fixerdate, "%Y-%m-%d")
-            .replace(tzinfo=dt.timezone.utc)
-            .date()
-        )
+        return dt.datetime.strptime(fixerdate, "%Y-%m-%d").astimezone().date()
     msg = f"Unknown date format {fixerdate!s} of type {type(fixerdate)!s} encountered"
     raise TypeError(
         msg,
@@ -212,7 +208,7 @@ def get_previous_business_day_before_today(
 
     """
     if today is None:
-        today = dt.datetime.now(tz=dt.timezone.utc).date()
+        today = dt.datetime.now().astimezone().date()
 
     return date_offset_foll(
         today - dt.timedelta(days=1),
@@ -241,8 +237,8 @@ def offset_business_days(
     ddate: datetime.date
         A starting date that does not have to be a business day
     days: int
-        The number of business days to offset from the business day that is
-        the closest preceding the day given
+        The number of business days to offset from the business day that is given
+        If days is set as anything other than an integer its value is set to zero
     countries: CountriesType, default: "SE"
         (List of) country code(s) according to ISO 3166-1 alpha-2
     custom_holidays: HolidayType, optional
@@ -255,6 +251,11 @@ def offset_business_days(
         The new offset business day
 
     """
+    try:
+        days = int(days)
+    except TypeError:
+        days = 0
+
     if days <= 0:
         scaledtoyeardays = int((days * 372 / 250) // 1) - 365
         ndate = ddate + dt.timedelta(days=scaledtoyeardays)

{openseries-1.5.6 → openseries-1.5.7}/openseries/frame.py

@@ -17,13 +17,17 @@ from numpy import (
     array,
     cov,
     cumprod,
+    divide,
     dot,
     float64,
     inf,
+    isinf,
     linspace,
     log,
     nan,
     sqrt,
+    square,
+    std,
     zeros,
 )
 from numpy import (
@@ -53,10 +57,6 @@ from statsmodels.regression.linear_model import (  # type: ignore[import-untyped
 from typing_extensions import Self
 
 from openseries._common_model import _CommonModel
-from openseries._risk import (
-    _calc_inv_vol_weights,
-    _ewma_calc,
-)
 from openseries.datefixer import do_resample_to_business_period_ends
 from openseries.load_plotly import load_plotly_dict
 from openseries.series import OpenTimeSeries
@@ -591,17 +591,13 @@ class OpenFrame(_CommonModel):
         raw_corr = [raw_cov[0] / (2 * raw_one[0] * raw_two[0])]
 
         for _, row in data.iloc[1:].iterrows():
-            tmp_raw_one = _ewma_calc(
-                reeturn=row.loc[cols[0], ValueType.RTRN],
-                prev_ewma=raw_one[-1],
-                time_factor=time_factor,
-                lmbda=lmbda,
+            tmp_raw_one = sqrt(
+                square(row.loc[cols[0], ValueType.RTRN]) * time_factor * (1 - lmbda)
+                + square(raw_one[-1]) * lmbda,
             )
-            tmp_raw_two = _ewma_calc(
-                reeturn=row.loc[cols[1], ValueType.RTRN],
-                prev_ewma=raw_two[-1],
-                time_factor=time_factor,
-                lmbda=lmbda,
+            tmp_raw_two = sqrt(
+                square(row.loc[cols[1], ValueType.RTRN]) * time_factor * (1 - lmbda)
+                + square(raw_two[-1]) * lmbda,
             )
             tmp_raw_cov = (
                 row.loc[cols[0], ValueType.RTRN]
@@ -1508,8 +1504,9 @@ class OpenFrame(_CommonModel):
             if weight_strat == "eq_weights":
                 self.weights = [1.0 / self.item_count] * self.item_count
             elif weight_strat == "inv_vol":
-                weight_calc = list(_calc_inv_vol_weights(returns=dframe))
-                self.weights = weight_calc
+                vol = divide(1.0, std(dframe, axis=0, ddof=1))
+                vol[isinf(vol)] = nan
+                self.weights = list(divide(vol, vol.sum()))
             else:
                 msg = "Weight strategy not implemented"
                 raise NotImplementedError(msg)
@@ -2188,7 +2185,7 @@ def sharpeplot(  # noqa: C901
                 xhoverformat=".2%",
                 yhoverformat=".2%",
                 hovertext=[point_frame.loc["text", col]],
-                hovertemplate=("Return %{y}<br>Vol %{x}%{hovertext}"),
+                hovertemplate="Return %{y}<br>Vol %{x}%{hovertext}",
                 hoverlabel_align="right",
                 marker={"size": 20, "color": clr},
                 mode=point_frame_mode,

{openseries-1.5.6 → openseries-1.5.7}/openseries/series.py

@@ -15,6 +15,7 @@ from numpy import (
     isnan,
     log,
     sqrt,
+    square,
 )
 from pandas import (
     DataFrame,
@@ -28,9 +29,6 @@ from pydantic import model_validator
 from typing_extensions import Self
 
 from openseries._common_model import _CommonModel
-from openseries._risk import (
-    _ewma_calc,
-)
 from openseries.datefixer import date_fix, do_resample_to_business_period_ends
 from openseries.types import (
     Countries,
@@ -675,13 +673,10 @@ class OpenTimeSeries(_CommonModel):
         ]
 
         for item in data.loc[:, cast(int, (self.label, ValueType.RTRN))].iloc[1:]:
-            previous = rawdata[-1]
+            prev = rawdata[-1]
             rawdata.append(
-                _ewma_calc(
-                    reeturn=cast(float, item),
-                    prev_ewma=previous,
-                    time_factor=time_factor,
-                    lmbda=lmbda,
+                sqrt(
+                    square(item) * time_factor * (1 - lmbda) + square(prev) * lmbda,
                 ),
             )
 

{openseries-1.5.6 → openseries-1.5.7}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "openseries"
-version = "1.5.6"
+version = "1.5.7"
 description = "Tools for analyzing financial timeseries."
 authors = ["Martin Karrin <martin.karrin@captor.se>"]
 repository = "https://github.com/CaptorAB/openseries"
@@ -34,8 +34,8 @@ keywords = [
 
 [tool.poetry.dependencies]
 python = ">=3.9,<3.13"
-holidays = ">=0.30,<0.50"
-numpy = ">=1.23.2,<=2.0.0"
+holidays = ">=0.30,<1.0"
+numpy = ">=1.23.2,<=3.0.0"
 openpyxl = ">=3.1.2,<4.0.0"
 pandas = ">=2.1.2,<3.0.0"
 plotly = ">=5.18.0,<6.0.0"
@@ -47,16 +47,16 @@ scipy = ">=1.11.4,<2.0.0"
 statsmodels = ">=0.14.0,<1.0.0"
 
 [tool.poetry.group.dev.dependencies]
-coverage = "^7.5.3"
+coverage = "^7.5.4"
 coverage-badge = "^1.1.1"
 mypy = "^1.10.0"
 pandas-stubs = "^2.2.2.240603"
 pre-commit = "^3.7.1"
-pytest = "^8.2.1"
-ruff = "^0.4.7"
-types-openpyxl = "^3.1.2.20240601"
+pytest = "^8.2.2"
+ruff = "^0.4.10"
+types-openpyxl = "^3.1.4.20240621"
 types-python-dateutil = "^2.9.0.20240316"
-types-requests = "^2.32.0.20240602"
+types-requests = "^2.32.0.20240622"
 
 [build-system]
 requires = ["poetry-core>=1.8.3"]
@@ -107,3 +107,9 @@ max-statements = 66
 
 [tool.ruff.lint.pyupgrade]
 keep-runtime-typing = true
+
+[tool.pytest.ini_options]
+filterwarnings = [
+    "error",
+    "ignore::DeprecationWarning:holidays.*"
+]