Merge pull request #100 from rstudio/statsmodels

implement statsmodels handler

Author: isabelizimm

.github/workflows/tests.yml

@@ -32,7 +32,7 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          python -m pip install -e .[dev,torch]
+          python -m pip install -e .[dev,torch,statsmodels]
       - name: Run Tests
         run: |
           pytest -m 'not rsc_test' --cov --cov-report xml

setup.cfg

@@ -48,3 +48,6 @@ dev =
 
 torch =
     torch
+
+statsmodels =
+    statsmodels

vetiver/__init__.py

@@ -14,6 +14,7 @@
 from .handlers.base import BaseHandler, create_handler, InvalidModelError  # noqa
 from .handlers.sklearn import SKLearnHandler  # noqa
 from .handlers.torch import TorchHandler  # noqa
+from .handlers.statsmodels import StatsmodelsHandler  # noqa
 from .rsconnect import deploy_rsconnect  # noqa
 from .monitor import compute_metrics, pin_metrics, plot_metrics, _rolling_df  # noqa
 

vetiver/handlers/base.py

@@ -13,7 +13,7 @@ class InvalidModelError(Exception):
 
     def __init__(
         self,
-        message="The `model` argument must be a scikit-learn or torch model.",
+        message="The `model` argument must be a supported or custom type.",
     ):
         self.message = message
         super().__init__(self.message)
@@ -47,9 +47,9 @@ def create_handler(model, ptype_data):
     """
 
     raise InvalidModelError(
-        "Model must be an sklearn or torch model, or a \
-        custom handler must be used. See the docs for more info on custom handlers. \
-        https://rstudio.github.io/vetiver-python/advancedusage/custom_handler.html"
+        "Model must be a supported model type, or a "
+        "custom handler must be used. See the docs for more info on custom handlers and"
+        "supported types https://rstudio.github.io/vetiver-python/"
     )
 
 
@@ -88,13 +88,13 @@ def create_meta(
         url: str = None,
         required_pkgs: list = [],
     ):
-        """Create metadata for sklearn model"""
+        """Create metadata for a model"""
         meta = _model_meta(user, version, url, required_pkgs)
 
         return meta
 
     def construct_ptype(self):
-        """Create data prototype for torch model
+        """Create data prototype for a model
 
         Parameters
         ----------

vetiver/handlers/statsmodels.py

@@ -0,0 +1,69 @@
+import pandas as pd
+
+from ..meta import _model_meta
+from .base import BaseHandler
+
+sm_exists = True
+try:
+    import statsmodels.api
+except ImportError:
+    sm_exists = False
+
+
+class StatsmodelsHandler(BaseHandler):
+    """Handler class for creating VetiverModels with statsmodels.
+
+    Parameters
+    ----------
+    model : statsmodels
+        a trained and fit statsmodels model
+    """
+
+    model_class = staticmethod(lambda: statsmodels.base.wrapper.ResultsWrapper)
+
+    def __init__(self, model, ptype_data):
+        super().__init__(model, ptype_data)
+
+    def describe(self):
+        """Create description for statsmodels model"""
+        desc = f"Statsmodels {self.model.__class__} model."
+        return desc
+
+    def create_meta(
+        user: list = None,
+        version: str = None,
+        url: str = None,
+        required_pkgs: list = [],
+    ):
+        """Create metadata for statsmodel"""
+        required_pkgs = required_pkgs + ["statsmodels"]
+        meta = _model_meta(user, version, url, required_pkgs)
+
+        return meta
+
+    def handler_predict(self, input_data, check_ptype):
+        """Generates method for /predict endpoint in VetiverAPI
+
+        The `handler_predict` function executes at each API call. Use this
+        function for calling `predict()` and any other tasks that must be executed
+        at each API call.
+
+        Parameters
+        ----------
+        input_data:
+            Test data
+
+        Returns
+        -------
+        prediction
+            Prediction from model
+        """
+        if sm_exists:
+            if isinstance(input_data, (list, pd.DataFrame)):
+                prediction = self.model.predict(input_data)
+            else:
+                prediction = self.model.predict([input_data])
+        else:
+            raise ImportError("Cannot import `statsmodels`")
+
+        return prediction

vetiver/server.py

@@ -82,8 +82,8 @@ async def prediction(
 
             @app.post("/predict")
             async def prediction(input_data: Request):
-
                 y = await input_data.json()
+
                 prediction = self.model.handler_predict(y, check_ptype=self.check_ptype)
 
                 return {"prediction": prediction.tolist()}

vetiver/tests/test_statsmodels.py

@@ -0,0 +1,67 @@
+import pytest
+
+sm = pytest.importorskip("statsmodels.api", reason="statsmodels library not installed")
+
+statsmodels = pytest.importorskip(
+    "statsmodels", reason="statsmodels library not installed"
+)
+
+import numpy as np  # noqa
+import pandas as pd  # noqa
+from fastapi.testclient import TestClient  # noqa
+
+import vetiver  # noqa
+
+
+@pytest.fixture
+def build_sm():
+
+    X, y = vetiver.get_mock_data()
+    glm = sm.GLM(y, X).fit()
+
+    v = vetiver.VetiverModel(glm, "glm", X)
+    return v
+
+
+def test_vetiver_build(build_sm):
+    api = vetiver.VetiverAPI(build_sm)
+    client = TestClient(api.app)
+    data = [{"B": 0, "C": 0, "D": 0}]
+
+    response = vetiver.predict(endpoint=client, data=data)
+
+    assert response.iloc[0, 0] == 0.0
+    assert len(response) == 1
+
+
+def test_batch(build_sm):
+    api = vetiver.VetiverAPI(build_sm)
+    client = TestClient(api.app)
+    data = pd.DataFrame(np.random.randint(0, 100, size=(100, 4)), columns=list("ABCD"))
+
+    response = vetiver.predict(endpoint=client, data=data)
+
+    assert len(response) == 100
+
+
+def test_no_ptype(build_sm):
+    api = vetiver.VetiverAPI(build_sm, check_ptype=False)
+    client = TestClient(api.app)
+    data = [0, 0, 0]
+
+    response = vetiver.predict(endpoint=client, data=data)
+
+    assert response.iloc[0, 0] == 0.0
+    assert len(response) == 1
+
+
+def test_serialize(build_sm):
+    import pins
+
+    board = pins.board_temp(allow_pickle_read=True)
+    vetiver.vetiver_pin_write(board=board, model=build_sm)
+    assert isinstance(
+        board.pin_read("glm"),
+        statsmodels.genmod.generalized_linear_model.GLMResultsWrapper,
+    )
+    board.pin_delete("glm")