Merge pull request #39 from bbcho/v0286

V0286

.devcontainer/devcontainer.json

@@ -0,0 +1,48 @@
+// For format details, see https://aka.ms/devcontainer.json. For config options, see the
+// README at: https://github.com/devcontainers/templates/tree/main/src/python
+{
+  "name": "Python 3",
+  // Or use a Dockerfile or Docker Compose file. More info: https://containers.dev/guide/dockerfile
+  // "image": "mcr.microsoft.com/devcontainers/python:1-3.10-bookworm",
+  "image": "bbcho/base:0.0.2",
+  "customizations": {
+    "vscode": {
+      "extensions": [
+          "ms-python.python",
+          "ms-toolsai.jupyter",
+          "reditorsupport.r"
+
+      ],
+      "settings": {
+          "python.testing.pytestArgs": [
+              "."
+          ],
+          "python.testing.unittestEnabled": false,
+              "python.testing.pytestEnabled": true,
+              "python.formatting.provider": "black",
+              "python.linting.mypyEnabled": true,
+              "python.linting.enabled": true
+        }
+    }
+  }
+  //"features": {
+  // "ghcr.io/rocker-org/devcontainer-features/r-apt:latest": {},
+  // "ghcr.io/rocker-org/devcontainer-features/quarto-cli:1": {}
+  //}
+
+  // Features to add to the dev container. More info: https://containers.dev/features.
+  // "features": {},
+
+  // Use 'forwardPorts' to make a list of ports inside the container available locally.
+  // "forwardPorts": [],
+
+  // Use 'postCreateCommand' to run commands after the container is created.
+  // "postCreateCommand": "pip3 install --user -r requirements.txt",
+
+  // Configure tool-specific properties.
+  // "customizations": {},
+
+  // Uncomment to connect as root instead. More info: https://aka.ms/dev-containers-non-root.
+  // "remoteUser": "root"
+}
+

.github/workflows/python-publish.yml

@@ -15,8 +15,8 @@ jobs:
   Linux-build:
     runs-on: ubuntu-latest
     env:
-      TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
-      TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
+      TWINE_USERNAME: __token__
+      TWINE_PASSWORD: ${{ secrets.PYPI_API_TOKEN }}
     steps:
       - uses: actions/checkout@v3
       - name: build manylinux wheels
@@ -37,8 +37,8 @@ jobs:
   Matrix-build:
     runs-on: ${{ matrix.os }}
     env:
-      TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
-      TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
+      TWINE_USERNAME: __token__
+      TWINE_PASSWORD: ${{ secrets.PYPI_API_TOKEN }}
     strategy:
       matrix:
         os: [windows-latest] #, macos-latest]
@@ -76,8 +76,8 @@ jobs:
         pip install setuptools wheel twine cython numpy
     - name: Build and publish
       env:
-        TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
-        TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
+        TWINE_USERNAME: __token__
+        TWINE_PASSWORD: ${{ secrets.PYPI_API_TOKEN }}
       run: |
         python setup.py sdist
         twine upload dist/*

src/risktools/_multivariate.py

@@ -152,7 +152,7 @@ def generate_eps_MV(cor, T, dt, sims=1000, mu=None, seed=None):
     return eps
 
 
-def simGBM_MV(s0, r, sigma, T, dt, mu=None, cor=None, eps=None, sims=1000):
+def simGBM_MV(s0, r, sigma, T, dt, mu=None, cor=None, eps=None, sims=1000, seed=None):
     """
     Simulate Geometric Brownian Motion for stochastic processes with
     multiple assets using a multivariate normal distribution.
@@ -187,6 +187,8 @@ def simGBM_MV(s0, r, sigma, T, dt, mu=None, cor=None, eps=None, sims=1000):
         simulations, and M is the number of assets. By default None.
     sims : int
         Number of simulations. By default 1000.
+    seed : int | None
+        To pass to numpy random number generator as seed. For testing only.
 
     Returns
     -------
@@ -220,7 +222,7 @@ def simGBM_MV(s0, r, sigma, T, dt, mu=None, cor=None, eps=None, sims=1000):
     N = int(T / dt)
 
     if eps is None:
-        eps = generate_eps_MV(cor, T, dt, sims, mu)
+        eps = generate_eps_MV(cor, T, dt, sims, mu, seed=seed)
 
     s = _np.zeros((N + 1, sims, len(s0)))
 
@@ -245,7 +247,7 @@ def simOU_MV(
     eps=None,
     seed=None,
     log_price=False,
-    **kwargs
+    **kwargs,
 ):
     """
     Simulate Ornstein-Uhlenbeck process for stochastic processes for
@@ -352,6 +354,16 @@ def simOU_MV(
 
     s = _np.zeros((N + 1, eps.shape[1], eps.shape[2]))
 
+    try:
+        s0 = s0.to_numpy()
+    except:
+        pass
+
+    try:
+        theta = theta.to_numpy()
+    except:
+        pass
+
     for i in range(0, eps.shape[2]):
         s[:, :, i] = simOU(
             s0=s0[i],
@@ -362,7 +374,7 @@ def simOU_MV(
             dt=dt,
             eps=eps[:, :, i],
             log_price=log_price,
-            **kwargs
+            **kwargs,
         )
 
     return s
@@ -385,7 +397,7 @@ def simOUJ_MV(
     elp=None,
     ejp=None,
     seed=None,
-    **kwargs
+    **kwargs,
 ):
     """
     Simulate Ornstein-Uhlenbeck Jump process for stochastic processes for
@@ -535,7 +547,7 @@ def simOUJ_MV(
             ejp=ejp_tmp,
             seed=seed,
             mr_lag=mr_lag[i],
-            **kwargs
+            **kwargs,
         )
 
     return s
@@ -1087,12 +1099,65 @@ def fit(self):
             self._sigma = returns.std() * _np.sqrt(1 / self._dt)
             self._cor = returns.corr()
 
-    def simulate(self, sims=1000):
+    def simulate(self, sims=1000, seed=None):
 
         self._sims = simGBM_MV(
-            self._s0, self._r, self._sigma, self._T, self._dt, cor=self._cor, sims=sims
+            self._s0,
+            self._r,
+            self._sigma,
+            self._T,
+            self._dt,
+            cor=self._cor,
+            sims=sims,
+            seed=seed,
         )
 
+    def output(self, start_date=None, freq="B", names=None):
+        """
+        Method for outputing the results of the simulations. It will return a
+        dictionary of dataframes where the names are the column names of the original
+        pricing dataframe passed to the object if it exists.
+
+        Parameters
+        ----------
+        start_date : str or datetime, optional
+            Start date of the simulation. By default None.
+        freq : str, optional
+            Frequency of the simulation. By default 'B'.
+        names : list[str], optional
+            List of strings to use as asset names in the output dictionary. Should be of length N
+            where N is the number of assets in the portfolio. By default None.
+
+        Returns
+        -------
+        Dictionary of dataframes where the keys are the column names of the original
+        pricing dataframe passed to the object if it exists.
+        """
+
+        df = self._sims.copy()
+
+        if (names is None) & (self._prices is not None):
+            names = self._prices.columns
+        else:
+            names = [f"Asset {str(i)}" for i in range(0, df.shape[2])]
+
+        if start_date is None:
+            start_date = _pd.Timestamp.now().floor("D") + _pd.Timedelta(days=1)
+
+        dates = _pd.date_range(start=start_date, periods=df.shape[0], freq=freq)
+
+        out = _pd.DataFrame()
+        for i, nm in enumerate(names):
+            tf = _pd.DataFrame(df[:, :, i], index=dates)
+            tf["asset"] = nm
+            tf.index.name = "date"
+            tf.columns.name = "sims"
+            tf = tf.reset_index().set_index(["asset", "date"])
+
+            out = _pd.concat([out, tf], axis=0)
+
+        return out
+
 
 class MVOU(_MVSIM):
     """
@@ -1196,12 +1261,17 @@ def fit(self, log_price=False, method="OLS", verbose=False, s0=None):
         # Only relevent if prices not passed to constructor
         if self._prices is not None:
             prices = _pd.DataFrame(self._prices)
-            returns = (_np.log(prices) - _np.log(prices.shift())).dropna()
+
+            # try:
+            # returns = (_np.log(prices) - _np.log(prices.shift())).dropna()
+            # except:
+            returns = prices.diff().dropna()
 
             self._s0 = prices.iloc[-1, :] if s0 is None else s0
+            self._s0 = _np.array(self._s0)
 
             self._params = fitOU_MV(
-                prices, self._dt, log_price=log_price, method=method, verbose=verbose
+                df=prices, dt=self._dt, log_price=log_price, method=method, verbose=verbose
             )
             self._cor = returns.corr()
         else:
@@ -1214,15 +1284,63 @@ def fit(self, log_price=False, method="OLS", verbose=False, s0=None):
         if self._asset_names is not None:
             self._params.columns = self._asset_names
 
-    def simulate(self, sims=1000):
+    def simulate(self, sims=1000, seed=None):
 
-        self._sims = self._sims = simOU_MV(
-            self._s0,
-            self._params.loc["mu", :],
-            self._params.loc["theta", :],
-            self._T,
-            self._dt,
-            self._params.loc["annualized_sigma", :],
-            self._cor,
+        self._sims = simOU_MV(
+            s0=self._s0,
+            mu=self._params.loc["mu", :],
+            theta=self._params.loc["theta", :],
+            T=self._T,
+            dt=self._dt,
+            sigma=self._params.loc["annualized_sigma", :],
+            cor=self._cor,
             sims=sims,
+            log_price=False,
+            seed=seed,
         )
+
+    def output(self, start_date=None, freq="B", names=None):
+        """
+        Method for outputing the results of the simulations. It will return a
+        dictionary of dataframes where the names are the column names of the original
+        pricing dataframe passed to the object if it exists.
+
+        Parameters
+        ----------
+        start_date : str or datetime, optional
+            Start date of the simulation. By default None.
+        freq : str, optional
+            Frequency of the simulation. By default 'B'.
+        names : list[str], optional
+            List of strings to use as asset names in the output dictionary. Should be of length N
+            where N is the number of assets in the portfolio. By default None.
+
+        Returns
+        -------
+        Dictionary of dataframes where the keys are the column names of the original
+        pricing dataframe passed to the object if it exists.
+        """
+
+        df = self._sims.copy()
+
+        if (names is None) & (self._prices is not None):
+            names = self._prices.columns
+        else:
+            names = [f"Asset {str(i)}" for i in range(0, df.shape[2])]
+
+        if start_date is None:
+            start_date = _pd.Timestamp.now().floor("D") + _pd.Timedelta(days=1)
+
+        dates = _pd.date_range(start=start_date, periods=df.shape[0], freq=freq)
+
+        out = _pd.DataFrame()
+        for i, nm in enumerate(names):
+            tf = _pd.DataFrame(df[:, :, i], index=dates)
+            tf["asset"] = nm
+            tf.index.name = "date"
+            tf.columns.name = "sims"
+            tf = tf.reset_index().set_index(["asset", "date"])
+
+            out = _pd.concat([out, tf], axis=0)
+
+        return out

src/risktools/_sims.py

@@ -31,15 +31,17 @@ def is_iterable(x):
 def make_into_array(x, N):
     # make an array of same size as N+1
     if is_iterable(x):
-        # if isinstance(x, _pd.DataFrame) == False:
-        #     x = _pd.DataFrame(x)
-        if len(x.shape) == 2:
+        if isinstance(x, _pd.DataFrame) == False:
+            x = _pd.DataFrame(x)
+        if x.shape[1] >= 2:
             # if a 2D array is passed, return it as is
             # good for stocastic volatility matrix
+            # but stack first row
             x = _np.vstack((x.iloc[0], x))
             return x
 
         if x.shape[0] == N:
+            # stack first row
             x = _np.append(x.iloc[0], x)
         else:
             raise ValueError(
@@ -48,6 +50,7 @@ def make_into_array(x, N):
 
     else:
         x = _np.ones(N + 1) * x
+
 
     return x
 
@@ -223,6 +226,7 @@ def simOU(
         mu = _np.tile(_np.array(mu), sims)
     else:
         mu = mu.flatten("F")
+
 
     # Don't run if 2D array passed for sigma
     if len(sigma.shape) == 1: