openseries 1.9.3__tar.gz → 1.9.4__tar.gz

{openseries-1.9.3 → openseries-1.9.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: openseries
-Version: 1.9.3
+Version: 1.9.4
 Summary: Tools for analyzing financial timeseries.
 License: # BSD 3-Clause License
          

{openseries-1.9.3 → openseries-1.9.4}/openseries/_common_model.py

@@ -25,7 +25,9 @@ from .owntypes import (
     DateAlignmentError,
     InitialValueZeroError,
     NumberOfItemsAndLabelsNotSameError,
+    ResampleDataLossError,
     Self,
+    ValueType,
 )
 
 if TYPE_CHECKING:  # pragma: no cover
@@ -47,7 +49,6 @@ if TYPE_CHECKING:  # pragma: no cover
         LiteralPlotlyJSlib,
         LiteralPlotlyOutput,
         LiteralQuantileInterp,
-        ValueType,
     )
 from openpyxl.utils.dataframe import dataframe_to_rows
 from openpyxl.workbook.workbook import Workbook
@@ -432,7 +433,17 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
 
         wmdf = wmdf.reindex(index=[deyt.date() for deyt in dates], method=method)
         wmdf.index = DatetimeIndex(wmdf.index)
-        result = wmdf.pct_change().min()
+
+        vtypes = [x == ValueType.RTRN for x in wmdf.columns.get_level_values(1)]
+        if any(vtypes):
+            msg = (
+                "Do not run worst_month on return series. The operation will "
+                "pick the last data point in the sparser series. It will not sum "
+                "returns and therefore data will be lost and result will be wrong."
+            )
+            raise ResampleDataLossError(msg)
+
+        result = wmdf.ffill().pct_change().min()
 
         if self.tsdf.shape[1] == 1:
             return float(result.iloc[0])
@@ -1305,6 +1316,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
 
         result = (
             self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()
             .mean()
             * time_factor
@@ -1366,7 +1378,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
             time_factor = how_many / fraction
 
         data = self.tsdf.loc[cast("int", earlier) : cast("int", later)]
-        result = data.pct_change().std().mul(sqrt(time_factor))
+        result = data.ffill().pct_change().std().mul(sqrt(time_factor))
 
         if self.tsdf.shape[1] == 1:
             return float(cast("SupportsFloat", result.iloc[0]))
@@ -1564,21 +1576,24 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         if drift_adjust:
             imp_vol = (-sqrt(time_factor) / norm.ppf(level)) * (
                 self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+                .ffill()
                 .pct_change()
                 .quantile(1 - level, interpolation=interpolation)
                 - self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+                .ffill()
                 .pct_change()
                 .sum()
                 / len(
-                    self.tsdf.loc[
-                        cast("int", earlier) : cast("int", later)
-                    ].pct_change(),
+                    self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+                    .ffill()
+                    .pct_change(),
                 )
             )
         else:
             imp_vol = (
                 -sqrt(time_factor)
                 * self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+                .ffill()
                 .pct_change()
                 .quantile(1 - level, interpolation=interpolation)
                 / norm.ppf(level)
@@ -1644,12 +1659,14 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         )
         result = [
             cvar_df.loc[:, x]  # type: ignore[call-overload,index]
+            .ffill()
             .pct_change()
             .sort_values()
             .iloc[
                 : ceil(
                     (1 - level)
                     * cvar_df.loc[:, x]  # type: ignore[index]
+                    .ffill()
                     .pct_change()
                     .count(),
                 ),
@@ -1718,6 +1735,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
 
         how_many = (
             self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()
             .count(numeric_only=True)
         )
@@ -1734,6 +1752,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         per_period_mar = min_accepted_return / time_factor
         diff = (
             self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()
             .sub(per_period_mar)
         )
@@ -1845,6 +1864,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         )
         result: NDArray[float64] = skew(
             a=self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()
             .to_numpy(),
             bias=True,
@@ -1892,7 +1912,11 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
             to_dt=to_date,
         )
         result: NDArray[float64] = kurtosis(
-            a=(self.tsdf.loc[cast("int", earlier) : cast("int", later)].pct_change()),
+            a=(
+                self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+                .ffill()
+                .pct_change()
+            ),
             fisher=True,
             bias=True,
             nan_policy="omit",
@@ -1988,16 +2012,18 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         )
         pos = (
             self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()[1:][
-                self.tsdf.loc[cast("int", earlier) : cast("int", later)].pct_change()[
-                    1:
-                ]
+                self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+                .ffill()
+                .pct_change()[1:]
                 > zero
             ]
             .count()
         )
         tot = (
             self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()
             .count()
         )
@@ -2184,7 +2210,11 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
             from_dt=from_date,
             to_dt=to_date,
         )
-        retdf = self.tsdf.loc[cast("int", earlier) : cast("int", later)].pct_change()
+        retdf = (
+            self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
+            .pct_change()
+        )
         pos = retdf[retdf > min_accepted_return].sub(min_accepted_return).sum()
         neg = retdf[retdf < min_accepted_return].sub(min_accepted_return).sum()
         ratio = pos / -neg
@@ -2272,7 +2302,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
             period = "-".join([str(year), str(month).zfill(2)])
         vrdf = self.tsdf.copy()
         vrdf.index = DatetimeIndex(vrdf.index)
-        resultdf = DataFrame(vrdf.pct_change())
+        resultdf = DataFrame(vrdf.ffill().pct_change())
         result = resultdf.loc[period] + 1
         cal_period = result.cumprod(axis="index").iloc[-1] - 1
         if self.tsdf.shape[1] == 1:
@@ -2324,6 +2354,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         )
         result = (
             self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()
             .quantile(1 - level, interpolation=interpolation)
         )
@@ -2372,6 +2403,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         )
         result = (
             self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
             .pct_change()
             .rolling(observations, min_periods=observations)
             .sum()
@@ -2418,9 +2450,11 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
             from_dt=from_date,
             to_dt=to_date,
         )
-        zscframe = self.tsdf.loc[
-            cast("int", earlier) : cast("int", later)
-        ].pct_change()
+        zscframe = (
+            self.tsdf.loc[cast("int", earlier) : cast("int", later)]
+            .ffill()
+            .pct_change()
+        )
         result = (zscframe.iloc[-1] - zscframe.mean()) / zscframe.std()
 
         if self.tsdf.shape[1] == 1:
@@ -2489,6 +2523,7 @@ class _CommonModel(BaseModel):  # type: ignore[misc]
         ret_label = cast("tuple[str]", self.tsdf.iloc[:, column].name)[0]
         retseries = (
             Series(self.tsdf.iloc[:, column])
+            .ffill()
             .pct_change()
             .rolling(observations, min_periods=observations)
             .sum()

{openseries-1.9.3 → openseries-1.9.4}/openseries/frame.py

@@ -57,6 +57,7 @@ from .owntypes import (
     NoWeightsError,
     OpenFramePropertiesList,
     RatioInputError,
+    ResampleDataLossError,
     Self,
     ValueType,
 )
@@ -340,7 +341,7 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             The returns of the values in the series
 
         """
-        returns = self.tsdf.pct_change()
+        returns = self.tsdf.ffill().pct_change()
         returns.iloc[0] = 0
         new_labels: list[ValueType] = [ValueType.RTRN] * self.item_count
         arrays: list[Index[Any], list[ValueType]] = [  # type: ignore[type-arg]
@@ -387,7 +388,7 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
         """
         vtypes = [x == ValueType.RTRN for x in self.tsdf.columns.get_level_values(1)]
         if not any(vtypes):
-            returns = self.tsdf.pct_change()
+            returns = self.tsdf.ffill().pct_change()
             returns.iloc[0] = 0
         elif all(vtypes):
             returns = self.tsdf.copy()
@@ -424,12 +425,27 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             An OpenFrame object
 
         """
+        vtypes = [x == ValueType.RTRN for x in self.tsdf.columns.get_level_values(1)]
+        if not any(vtypes):
+            value_type = ValueType.PRICE
+        elif all(vtypes):
+            value_type = ValueType.RTRN
+        else:
+            msg = "Mix of series types will give inconsistent results"
+            raise MixedValuetypesError(msg)
+
         self.tsdf.index = DatetimeIndex(self.tsdf.index)
-        self.tsdf = self.tsdf.resample(freq).last()
+        if value_type == ValueType.PRICE:
+            self.tsdf = self.tsdf.resample(freq).last()
+        else:
+            self.tsdf = self.tsdf.resample(freq).sum()
         self.tsdf.index = Index(d.date() for d in DatetimeIndex(self.tsdf.index))
         for xerie in self.constituents:
             xerie.tsdf.index = DatetimeIndex(xerie.tsdf.index)
-            xerie.tsdf = xerie.tsdf.resample(freq).last()
+            if value_type == ValueType.PRICE:
+                xerie.tsdf = xerie.tsdf.resample(freq).last()
+            else:
+                xerie.tsdf = xerie.tsdf.resample(freq).sum()
             xerie.tsdf.index = Index(
                 dejt.date() for dejt in DatetimeIndex(xerie.tsdf.index)
             )
@@ -458,6 +474,15 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             An OpenFrame object
 
         """
+        vtypes = [x == ValueType.RTRN for x in self.tsdf.columns.get_level_values(1)]
+        if any(vtypes):
+            msg = (
+                "Do not run resample_to_business_period_ends on return series. "
+                "The operation will pick the last data point in the sparser series. "
+                "It will not sum returns and therefore data will be lost."
+            )
+            raise ResampleDataLossError(msg)
+
         for xerie in self.constituents:
             dates = _do_resample_to_business_period_ends(
                 data=xerie.tsdf,
@@ -530,7 +555,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             Series volatilities and correlation
 
         """
-        earlier, later = self.calc_range(months_from_last, from_date, to_date)
+        earlier, later = self.calc_range(
+            months_offset=months_from_last, from_dt=from_date, to_dt=to_date
+        )
         if periods_in_a_year_fixed is None:
             fraction = (later - earlier).days / 365.25
             how_many = (
@@ -621,9 +648,13 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             Correlation matrix
 
         """
-        corr_matrix = self.tsdf.pct_change().corr(
-            method="pearson",
-            min_periods=1,
+        corr_matrix = (
+            self.tsdf.ffill()
+            .pct_change()
+            .corr(
+                method="pearson",
+                min_periods=1,
+            )
         )
         corr_matrix.columns = corr_matrix.columns.droplevel(level=1)
         corr_matrix.index = corr_matrix.index.droplevel(level=1)
@@ -805,7 +836,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             Tracking Errors
 
         """
-        earlier, later = self.calc_range(months_from_last, from_date, to_date)
+        earlier, later = self.calc_range(
+            months_offset=months_from_last, from_dt=from_date, to_dt=to_date
+        )
         fraction = (later - earlier).days / 365.25
 
         msg = "base_column should be a tuple[str, ValueType] or an integer."
@@ -848,10 +881,8 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                     :,
                     item,
                 ]
-                relative = 1.0 + longdf - shortdf
-                vol = float(
-                    relative.pct_change().std() * sqrt(time_factor),
-                )
+                relative = longdf.ffill().pct_change() - shortdf.ffill().pct_change()
+                vol = float(relative.std() * sqrt(time_factor))
                 terrors.append(vol)
 
         return Series(
@@ -897,7 +928,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             Information Ratios
 
         """
-        earlier, later = self.calc_range(months_from_last, from_date, to_date)
+        earlier, later = self.calc_range(
+            months_offset=months_from_last, from_dt=from_date, to_dt=to_date
+        )
         fraction = (later - earlier).days / 365.25
 
         msg = "base_column should be a tuple[str, ValueType] or an integer."
@@ -940,13 +973,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                     :,
                     item,
                 ]
-                relative = 1.0 + longdf - shortdf
-                ret = float(
-                    relative.pct_change().mean() * time_factor,
-                )
-                vol = float(
-                    relative.pct_change().std() * sqrt(time_factor),
-                )
+                relative = longdf.ffill().pct_change() - shortdf.ffill().pct_change()
+                ret = float(relative.mean() * time_factor)
+                vol = float(relative.std() * sqrt(time_factor))
                 ratios.append(ret / vol)
 
         return Series(
@@ -1000,7 +1029,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
 
         """
         loss_limit: float = 0.0
-        earlier, later = self.calc_range(months_from_last, from_date, to_date)
+        earlier, later = self.calc_range(
+            months_offset=months_from_last, from_dt=from_date, to_dt=to_date
+        )
         fraction = (later - earlier).days / 365.25
 
         msg = "base_column should be a tuple[str, ValueType] or an integer."
@@ -1046,16 +1077,18 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                 msg = "ratio must be one of 'up', 'down' or 'both'."
                 if ratio == "up":
                     uparray = (
-                        longdf.pct_change()[
-                            shortdf.pct_change().to_numpy() > loss_limit
+                        longdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() > loss_limit
                         ]
                         .add(1)
                         .to_numpy()
                     )
                     up_rtrn = uparray.prod() ** (1 / (len(uparray) / time_factor)) - 1
                     upidxarray = (
-                        shortdf.pct_change()[
-                            shortdf.pct_change().to_numpy() > loss_limit
+                        shortdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() > loss_limit
                         ]
                         .add(1)
                         .to_numpy()
@@ -1066,8 +1099,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                     ratios.append(up_rtrn / up_idx_return)
                 elif ratio == "down":
                     downarray = (
-                        longdf.pct_change()[
-                            shortdf.pct_change().to_numpy() < loss_limit
+                        longdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() < loss_limit
                         ]
                         .add(1)
                         .to_numpy()
@@ -1076,8 +1110,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                         downarray.prod() ** (1 / (len(downarray) / time_factor)) - 1
                     )
                     downidxarray = (
-                        shortdf.pct_change()[
-                            shortdf.pct_change().to_numpy() < loss_limit
+                        shortdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() < loss_limit
                         ]
                         .add(1)
                         .to_numpy()
@@ -1089,16 +1124,18 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                     ratios.append(down_return / down_idx_return)
                 elif ratio == "both":
                     uparray = (
-                        longdf.pct_change()[
-                            shortdf.pct_change().to_numpy() > loss_limit
+                        longdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() > loss_limit
                         ]
                         .add(1)
                         .to_numpy()
                     )
                     up_rtrn = uparray.prod() ** (1 / (len(uparray) / time_factor)) - 1
                     upidxarray = (
-                        shortdf.pct_change()[
-                            shortdf.pct_change().to_numpy() > loss_limit
+                        shortdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() > loss_limit
                         ]
                         .add(1)
                         .to_numpy()
@@ -1107,8 +1144,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                         upidxarray.prod() ** (1 / (len(upidxarray) / time_factor)) - 1
                     )
                     downarray = (
-                        longdf.pct_change()[
-                            shortdf.pct_change().to_numpy() < loss_limit
+                        longdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() < loss_limit
                         ]
                         .add(1)
                         .to_numpy()
@@ -1117,8 +1155,9 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                         downarray.prod() ** (1 / (len(downarray) / time_factor)) - 1
                     )
                     downidxarray = (
-                        shortdf.pct_change()[
-                            shortdf.pct_change().to_numpy() < loss_limit
+                        shortdf.ffill()
+                        .pct_change()[
+                            shortdf.ffill().pct_change().to_numpy() < loss_limit
                         ]
                         .add(1)
                         .to_numpy()
@@ -1173,10 +1212,8 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             Beta as Co-variance of x & y divided by Variance of x
 
         """
-        if all(
-            x_value == ValueType.RTRN
-            for x_value in self.tsdf.columns.get_level_values(1).to_numpy()
-        ):
+        vtypes = [x == ValueType.RTRN for x in self.tsdf.columns.get_level_values(1)]
+        if all(vtypes):
             msg = "asset should be a tuple[str, ValueType] or an integer."
             if isinstance(asset, tuple):
                 y_value = self.tsdf.loc[:, asset]
@@ -1192,33 +1229,27 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
                 x_value = self.tsdf.iloc[:, market]
             else:
                 raise TypeError(msg)
-        else:
+        elif not any(vtypes):
             msg = "asset should be a tuple[str, ValueType] or an integer."
             if isinstance(asset, tuple):
-                y_value = log(
-                    self.tsdf.loc[:, asset] / self.tsdf.loc[:, asset].iloc[0],
-                )
+                y_value = self.tsdf.loc[:, asset].ffill().pct_change().iloc[1:]
             elif isinstance(asset, int):
-                y_value = log(
-                    self.tsdf.iloc[:, asset] / cast("float", self.tsdf.iloc[0, asset]),
-                )
+                y_value = self.tsdf.iloc[:, asset].ffill().pct_change().iloc[1:]
             else:
                 raise TypeError(msg)
-
             msg = "market should be a tuple[str, ValueType] or an integer."
+
             if isinstance(market, tuple):
-                x_value = log(
-                    self.tsdf.loc[:, market] / self.tsdf.loc[:, market].iloc[0],
-                )
+                x_value = self.tsdf.loc[:, market].ffill().pct_change().iloc[1:]
             elif isinstance(market, int):
-                x_value = log(
-                    self.tsdf.iloc[:, market]
-                    / cast("float", self.tsdf.iloc[0, market]),
-                )
+                x_value = self.tsdf.iloc[:, market].ffill().pct_change().iloc[1:]
             else:
                 raise TypeError(msg)
+        else:
+            msg = "Mix of series types will give inconsistent results"
+            raise MixedValuetypesError(msg)
 
-        covariance = cov(y_value, x_value, ddof=dlta_degr_freedms)
+        covariance = cov(m=y_value, y=x_value, ddof=dlta_degr_freedms)
         beta = covariance[0, 1] / covariance[1, 1]
 
         return float(beta)
@@ -1319,105 +1350,57 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
             Jensen's alpha
 
         """
-        full_year = 1.0
         vtypes = [x == ValueType.RTRN for x in self.tsdf.columns.get_level_values(1)]
         if not any(vtypes):
             msg = "asset should be a tuple[str, ValueType] or an integer."
             if isinstance(asset, tuple):
-                asset_log = log(
-                    self.tsdf.loc[:, asset] / self.tsdf.loc[:, asset].iloc[0],
-                )
-                if self.yearfrac > full_year:
-                    asset_cagr = (
-                        self.tsdf.loc[:, asset].iloc[-1]
-                        / self.tsdf.loc[:, asset].iloc[0]
-                    ) ** (1 / self.yearfrac) - 1
-                else:
-                    asset_cagr = (
-                        self.tsdf.loc[:, asset].iloc[-1]
-                        / self.tsdf.loc[:, asset].iloc[0]
-                        - 1
-                    )
+                asset_rtn = self.tsdf.loc[:, asset].ffill().pct_change().iloc[1:]
+                asset_rtn_mean = float(asset_rtn.mean() * self.periods_in_a_year)
             elif isinstance(asset, int):
-                asset_log = log(
-                    self.tsdf.iloc[:, asset] / cast("float", self.tsdf.iloc[0, asset]),
-                )
-                if self.yearfrac > full_year:
-                    asset_cagr = (
-                        cast("float", self.tsdf.iloc[-1, asset])
-                        / cast("float", self.tsdf.iloc[0, asset])
-                    ) ** (1 / self.yearfrac) - 1
-                else:
-                    asset_cagr = (
-                        cast("float", self.tsdf.iloc[-1, asset])
-                        / cast("float", self.tsdf.iloc[0, asset])
-                        - 1
-                    )
+                asset_rtn = self.tsdf.iloc[:, asset].ffill().pct_change().iloc[1:]
+                asset_rtn_mean = float(asset_rtn.mean() * self.periods_in_a_year)
             else:
                 raise TypeError(msg)
 
             msg = "market should be a tuple[str, ValueType] or an integer."
             if isinstance(market, tuple):
-                market_log = log(
-                    self.tsdf.loc[:, market] / self.tsdf.loc[:, market].iloc[0],
-                )
-                if self.yearfrac > full_year:
-                    market_cagr = (
-                        self.tsdf.loc[:, market].iloc[-1]
-                        / self.tsdf.loc[:, market].iloc[0]
-                    ) ** (1 / self.yearfrac) - 1
-                else:
-                    market_cagr = (
-                        self.tsdf.loc[:, market].iloc[-1]
-                        / self.tsdf.loc[:, market].iloc[0]
-                        - 1
-                    )
+                market_rtn = self.tsdf.loc[:, market].ffill().pct_change().iloc[1:]
+                market_rtn_mean = float(market_rtn.mean() * self.periods_in_a_year)
             elif isinstance(market, int):
-                market_log = log(
-                    self.tsdf.iloc[:, market]
-                    / cast("float", self.tsdf.iloc[0, market]),
-                )
-                if self.yearfrac > full_year:
-                    market_cagr = (
-                        cast("float", self.tsdf.iloc[-1, market])
-                        / cast("float", self.tsdf.iloc[0, market])
-                    ) ** (1 / self.yearfrac) - 1
-                else:
-                    market_cagr = (
-                        cast("float", self.tsdf.iloc[-1, market])
-                        / cast("float", self.tsdf.iloc[0, market])
-                        - 1
-                    )
+                market_rtn = self.tsdf.iloc[:, market].ffill().pct_change().iloc[1:]
+                market_rtn_mean = float(market_rtn.mean() * self.periods_in_a_year)
             else:
                 raise TypeError(msg)
         elif all(vtypes):
             msg = "asset should be a tuple[str, ValueType] or an integer."
             if isinstance(asset, tuple):
-                asset_log = self.tsdf.loc[:, asset]
-                asset_cagr = asset_log.mean()
+                asset_rtn = self.tsdf.loc[:, asset]
+                asset_rtn_mean = float(asset_rtn.mean() * self.periods_in_a_year)
             elif isinstance(asset, int):
-                asset_log = self.tsdf.iloc[:, asset]
-                asset_cagr = asset_log.mean()
+                asset_rtn = self.tsdf.iloc[:, asset]
+                asset_rtn_mean = float(asset_rtn.mean() * self.periods_in_a_year)
             else:
                 raise TypeError(msg)
 
             msg = "market should be a tuple[str, ValueType] or an integer."
             if isinstance(market, tuple):
-                market_log = self.tsdf.loc[:, market]
-                market_cagr = market_log.mean()
+                market_rtn = self.tsdf.loc[:, market]
+                market_rtn_mean = float(market_rtn.mean() * self.periods_in_a_year)
             elif isinstance(market, int):
-                market_log = self.tsdf.iloc[:, market]
-                market_cagr = market_log.mean()
+                market_rtn = self.tsdf.iloc[:, market]
+                market_rtn_mean = float(market_rtn.mean() * self.periods_in_a_year)
             else:
                 raise TypeError(msg)
         else:
             msg = "Mix of series types will give inconsistent results"
             raise MixedValuetypesError(msg)
 
-        covariance = cov(m=asset_log, y=market_log, ddof=dlta_degr_freedms)
+        covariance = cov(m=asset_rtn, y=market_rtn, ddof=dlta_degr_freedms)
         beta = covariance[0, 1] / covariance[1, 1]
 
-        return float(asset_cagr - riskfree_rate - beta * (market_cagr - riskfree_rate))
+        return float(
+            asset_rtn_mean - riskfree_rate - beta * (market_rtn_mean - riskfree_rate)
+        )
 
     def make_portfolio(
         self: Self,
@@ -1448,7 +1431,7 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
 
         vtypes = [x == ValueType.RTRN for x in self.tsdf.columns.get_level_values(1)]
         if not any(vtypes):
-            returns = self.tsdf.pct_change()
+            returns = self.tsdf.ffill().pct_change()
             returns.iloc[0] = 0
         elif all(vtypes):
             returns = self.tsdf.copy()
@@ -1523,11 +1506,14 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
         )
 
         retseries = (
-            relative.pct_change().rolling(observations, min_periods=observations).sum()
+            relative.ffill()
+            .pct_change()
+            .rolling(observations, min_periods=observations)
+            .sum()
         )
         retdf = retseries.dropna().to_frame()
 
-        voldf = relative.pct_change().rolling(
+        voldf = relative.ffill().pct_change().rolling(
             observations,
             min_periods=observations,
         ).std() * sqrt(time_factor)
@@ -1573,9 +1559,13 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
         asset_label = cast("tuple[str, str]", self.tsdf.iloc[:, asset_column].name)[0]
         beta_label = f"{asset_label} / {market_label}"
 
-        rolling = self.tsdf.pct_change().rolling(
-            observations,
-            min_periods=observations,
+        rolling = (
+            self.tsdf.ffill()
+            .pct_change()
+            .rolling(
+                observations,
+                min_periods=observations,
+            )
         )
 
         rcov = rolling.cov(ddof=dlta_degr_freedms)
@@ -1630,10 +1620,11 @@ class OpenFrame(_CommonModel):  # type: ignore[misc]
         )
         first_series = (
             self.tsdf.iloc[:, first_column]
+            .ffill()
             .pct_change()[1:]
             .rolling(observations, min_periods=observations)
         )
-        second_series = self.tsdf.iloc[:, second_column].pct_change()[1:]
+        second_series = self.tsdf.iloc[:, second_column].ffill().pct_change()[1:]
         corrdf = first_series.corr(other=second_series).dropna().to_frame()
         corrdf.columns = MultiIndex.from_arrays(
             [

{openseries-1.9.3 → openseries-1.9.4}/openseries/owntypes.py

@@ -376,3 +376,7 @@ class IncorrectArgumentComboError(Exception):
 
 class PropertiesInputValidationError(Exception):
     """Raised when duplicate strings are provided."""
+
+
+class ResampleDataLossError(Exception):
+    """Raised when user attempts to run resample_to_business_period_ends on returns."""

{openseries-1.9.3 → openseries-1.9.4}/openseries/portfoliotools.py

@@ -320,7 +320,7 @@ def efficient_frontier(
 
     if tweak:
         limit_tweak = 0.001
-        line_df["stdev_diff"] = line_df.stdev.pct_change()
+        line_df["stdev_diff"] = line_df.stdev.ffill().pct_change()
         line_df = line_df.loc[line_df.stdev_diff.abs() > limit_tweak]
         line_df = line_df.drop(columns="stdev_diff")
 

{openseries-1.9.3 → openseries-1.9.4}/openseries/report.py

@@ -68,9 +68,7 @@ def calendar_period_returns(
     """
     copied = data.from_deepcopy()
     copied.resample_to_business_period_ends(freq=freq)
-    vtypes = [x == ValueType.RTRN for x in copied.tsdf.columns.get_level_values(1)]
-    if not any(vtypes):
-        copied.value_to_ret()
+    copied.value_to_ret()
     cldr = copied.tsdf.iloc[1:].copy()
     if relabel:
         if freq.upper() == "BYE":

{openseries-1.9.3 → openseries-1.9.4}/openseries/series.py

@@ -53,6 +53,7 @@ from .owntypes import (
     LiteralSeriesProps,
     MarketsNotStringNorListStrError,
     OpenTimeSeriesPropertiesList,
+    ResampleDataLossError,
     Self,
     ValueListType,
     ValueType,
@@ -66,7 +67,7 @@ TypeOpenTimeSeries = TypeVar("TypeOpenTimeSeries", bound="OpenTimeSeries")
 
 
 # noinspection PyUnresolvedReferences,PyNestedDecorators
-class OpenTimeSeries(_CommonModel):
+class OpenTimeSeries(_CommonModel):  # type: ignore[misc]
     """OpenTimeSeries objects are at the core of the openseries package.
 
     The intended use is to allow analyses of financial timeseries.
@@ -468,7 +469,7 @@ class OpenTimeSeries(_CommonModel):
             The returns of the values in the series
 
         """
-        returns = self.tsdf.pct_change()
+        returns = self.tsdf.ffill().pct_change()
         returns.iloc[0] = 0
         self.valuetype = ValueType.RTRN
         arrays = [[self.label], [self.valuetype]]
@@ -586,7 +587,10 @@ class OpenTimeSeries(_CommonModel):
 
         """
         self.tsdf.index = DatetimeIndex(self.tsdf.index)
-        self.tsdf = self.tsdf.resample(freq).last()
+        if self.valuetype == ValueType.RTRN:
+            self.tsdf = self.tsdf.resample(freq).sum()
+        else:
+            self.tsdf = self.tsdf.resample(freq).last()
         self.tsdf.index = Index(d.date() for d in DatetimeIndex(self.tsdf.index))
         return self
 
@@ -612,6 +616,14 @@ class OpenTimeSeries(_CommonModel):
             An OpenTimeSeries object
 
         """
+        if self.valuetype == ValueType.RTRN:
+            msg = (
+                "Do not run resample_to_business_period_ends on return series. "
+                "The operation will pick the last data point in the sparser series. "
+                "It will not sum returns and therefore data will be lost."
+            )
+            raise ResampleDataLossError(msg)
+
         dates = _do_resample_to_business_period_ends(
             data=self.tsdf,
             freq=freq,
@@ -659,7 +671,9 @@ class OpenTimeSeries(_CommonModel):
             Series EWMA volatility
 
         """
-        earlier, later = self.calc_range(months_from_last, from_date, to_date)
+        earlier, later = self.calc_range(
+            months_offset=months_from_last, from_dt=from_date, to_dt=to_date
+        )
         if periods_in_a_year_fixed:
             time_factor = float(periods_in_a_year_fixed)
         else:
@@ -725,7 +739,7 @@ class OpenTimeSeries(_CommonModel):
             returns_input = True
         else:
             values = [cast("float", self.tsdf.iloc[0, 0])]
-            ra_df = self.tsdf.pct_change()
+            ra_df = self.tsdf.ffill().pct_change()
             returns_input = False
         ra_df = ra_df.dropna()
 

{openseries-1.9.3 → openseries-1.9.4}/openseries/simulation.py

@@ -129,7 +129,9 @@ class ReturnSimulation(BaseModel):  # type: ignore[misc]
         """
         return cast(
             "float",
-            (self.results.pct_change().mean() * self.trading_days_in_year).iloc[0],
+            (
+                self.results.ffill().pct_change().mean() * self.trading_days_in_year
+            ).iloc[0],
         )
 
     @property
@@ -144,9 +146,10 @@ class ReturnSimulation(BaseModel):  # type: ignore[misc]
         """
         return cast(
             "float",
-            (self.results.pct_change().std() * sqrt(self.trading_days_in_year)).iloc[
-                0
-            ],
+            (
+                self.results.ffill().pct_change().std()
+                * sqrt(self.trading_days_in_year)
+            ).iloc[0],
         )
 
     @classmethod

{openseries-1.9.3 → openseries-1.9.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "openseries"
-version = "1.9.3"
+version = "1.9.4"
 description = "Tools for analyzing financial timeseries."
 authors = [
     { name = "Martin Karrin", email = "martin.karrin@captor.se" },
@@ -67,7 +67,7 @@ pre-commit = ">=3.7.1,<6.0.0"
 pytest = ">=8.2.2,<9.0.0"
 pytest-cov = ">=5.0.0,<7.0.0"
 pytest-xdist = ">=3.3.1,<5.0.0"
-ruff = "0.12.8"
+ruff = "0.12.10"
 types-openpyxl = ">=3.1.2,<5.0.0"
 types-python-dateutil = ">=2.8.2,<4.0.0"
 types-requests = ">=2.20.0,<3.0.0"