Merge branch 'main' into handlers-update

Author: isabelizimm

MANIFEST.in

@@ -1 +0,0 @@
-include vetiver/data/*.csv

docs/source/conf.py

@@ -1,6 +1,7 @@
 import os
 import sys
-from vetiver import __version__
+import re
+from vetiver import __version__ as version
 
 sys.path.insert(0, os.path.abspath("."))
 
@@ -12,8 +13,14 @@
 author = "Isabel Zimmerman"
 
 # The full version, including alpha/beta/rc tags
-release = __version__
-
+# Do not use the 0.0 version created by setuptools_scm when the package
+# is cloned too shallow to pickup
+p = re.compile(r"^0\.0\.post\d+\+g")
+if p.match(version):
+    commit = p.sub("", version)
+    version = f"Commit: {commit}"
+
+release = version
 
 # -- General configuration ---------------------------------------------------
 

pyproject.toml

@@ -1,5 +1,9 @@
 [build-system]
-requires = ["setuptools>=45", "wheel"]
+requires = [
+    "setuptools>=59",
+    "setuptools_scm[toml]>=6.4",
+    "wheel"
+]
 build-backend = "setuptools.build_meta"
 
 [tool.pytest.ini_options]
@@ -9,3 +13,7 @@ doctest_optionflags = "NORMALIZE_WHITESPACE"
 markers = [
     "rsc_test: tests for rstudio connect",
 ]
+
+[tool.setuptools_scm]
+fallback_version = "999"
+version_scheme = 'post-release'

setup.cfg

@@ -35,6 +35,7 @@ install_requires =
     pins
     rsconnect-python
     plotly
+    importlib-metadata>=4.4  # NOTE: Remove when python_requires>=3.8
 
 [options.extras_require]
 dev =

setup.py

@@ -1,17 +1,4 @@
-import io
-import os
-import re
-
-from setuptools import find_packages
 from setuptools import setup
 
 
-def read(filename):
-    filename = os.path.join(os.path.dirname(__file__), filename)
-    text_type = type("")
-    with io.open(filename, mode="r", encoding="utf-8") as fd:
-        return re.sub(text_type(r":[a-z]+:`~?(.*?)`"), text_type(r"``\1``"), fd.read())
-
-
-if __name__ == "__main__":
-    setup()
+setup()

vetiver/__init__.py

@@ -1,10 +1,7 @@
 """vetiver - Python parallel to R vetiver package"""
+# Change to import.metadata when minimum python>=3.8
+from importlib_metadata import version as _version
 
-__version__ = "0.1.5"
-__author__ = "Isabel Zimmerman <isabel.zimmerman@rstudio.com>"
-__all__ = []
-
-import importlib  # noqa
 from .ptype import *  # noqa
 from .vetiver_model import VetiverModel  # noqa
 from .server import VetiverAPI, vetiver_endpoint, predict  # noqa
@@ -17,5 +14,10 @@
 from .handlers.base import BaseHandler, create_handler, InvalidModelError  # noqa
 from .handlers.sklearn import SKLearnHandler  # noqa
 from .handlers.torch import TorchHandler  # noqa
-from .rsconnect import deploy_rsconnect  # noqa
-from .monitor import compute_metrics, pin_metrics, plot_metrics, _rolling_df  # noqa
+from .rsconnect import deploy_rsconnect # noqa
+from .monitor import compute_metrics, pin_metrics, plot_metrics, _rolling_df # noqa
+
+__author__ = "Isabel Zimmerman <isabel.zimmerman@rstudio.com>"
+__all__ = []
+__version__ = _version("vetiver")
+del _version

vetiver/monitor.py

@@ -1,9 +1,6 @@
-import datetime
-import pins
-from pins.errors import PinsError
 import plotly.express as px
 import pandas as pd
-from datetime import datetime, timedelta
+from datetime import timedelta
 
 
 def compute_metrics(
@@ -75,60 +72,74 @@ def _rolling_df(df: pd.DataFrame, td: timedelta):
         first = stop
 
 
-def pin_metrics(board, df_metrics, metrics_pin_name, overwrite=False):
-    pass
+def pin_metrics(
+    board,
+    df_metrics: pd.DataFrame,
+    metrics_pin_name: str,
+    pin_type: "str | None" = None,
+    index_name: str = "index",
+    overwrite: bool = False,
+) -> pd.DataFrame:
+    """
+    Update an existing pin storing model metrics over time
+
+    Parameters
+    ----------
+    board :
+        Pins board
+    df_metrics: pd.DataFrame
+        Dataframe of metrics over time, such as created by `vetiver_compute_metrics()`
+    metrics_pin_name:
+        Pin name for where the metrics are stored
+    index_name:
+        The column in df_metrics containing the aggregated dates or datetimes.
+        Note that this defaults to a column named "index".
+    overwrite: bool
+        If TRUE (the default), overwrite any metrics for
+        dates that exist both in the existing pin and
+        new metrics with the new values. If FALSE, error
+        when the new metrics contain overlapping dates with
+        the existing pin.
+    """
 
+    old_metrics_raw = board.pin_read(metrics_pin_name)
 
-#     """
-#     Update an existing pin storing model metrics over time
+    # need to coerce date index to a datetime, since pandas does not infer
+    # date columns from CSV (but note that formats like arrow do)
+    old_metrics = old_metrics_raw.copy()
+    old_metrics[index_name] = pd.to_datetime(old_metrics[index_name])
 
-#     Parameters
-#     ----------
-#     board :
-#         Pins board
-#     df_metrics: pd.DataFrame
-#         Dataframe of metrics over time, such as created by `vetiver_compute_metrics()`
-#     metrics_pin_name:
-#         Pin name for where the metrics are stored
-#     overwrite: bool
-#         If TRUE (the default), overwrite any metrics for
-#         dates that exist both in the existing pin and
-#         new metrics with the new values. If FALSE, error
-#         when the new metrics contain overlapping dates with
-#         the existing pin.
-#     """
-#     date_types = (datetime.date, datetime.time, datetime.datetime)
-#     if not isinstance(df_metrics.index, date_types):
-#         try:
-#             df_metrics = df_metrics.index.astype("datetime")
-#         except TypeError:
-#             raise TypeError(f"Index of {df_metrics} must be a date type")
+    # handle overlapping dates ----
+    dt_new = pd.to_datetime(df_metrics[index_name])
+    dt_old = old_metrics[index_name]
 
-#     new_metrics = df_metrics.sort_index()
+    indx_old_overlap = dt_old.isin(dt_new)
 
-#     new_dates = df_metrics.index.unique()
+    if overwrite:
+        # get only rows specific to old metrics, so when we concat below
+        # it effectively is an upsert
+        old_metrics = old_metrics.loc[~indx_old_overlap, :]
 
-#     try:
-#         old_metrics = board.pin_read(metrics_pin_name)
-#     except PinsError:
-#         board.pin_write(metrics_pin_name)
+    elif not overwrite and indx_old_overlap.any():
+        raise ValueError(
+            f"The new metrics overlap with dates already stored in {metrics_pin_name}."
+            " Check the aggregated dates or use `overwrite=True`."
+        )
 
-#     overlapping_dates = old_metrics.index in new_dates
+    # update and pin ----
+    combined_metrics = pd.concat([old_metrics, df_metrics], ignore_index=True)
+    sorted_metrics = combined_metrics.sort_values(index_name)
 
-#     if overwrite is True:
-#         old_metrics = old_metrics not in overlapping_dates
-#     else:
-#         if overlapping_dates:
-#             raise ValueError(
-#                 f"The new metrics overlap with dates \
-#                      already stored in {repr(metrics_pin_name)} \
-#                      Check the aggregated dates or use `overwrite = True`"
-#             )
+    if pin_type is None:
+        meta = board.pin_meta(metrics_pin_name)
 
-#     new_metrics = old_metrics + df_metrics
-#     new_metrics = new_metrics.sort_index()
+        final_pin_type = meta.type
+    else:
+        final_pin_type = pin_type
 
-#     pins.pin_write(board, new_metrics, metrics_pin_name)
+    board.pin_write(sorted_metrics, metrics_pin_name, type=final_pin_type)
+
+    return sorted_metrics
 
 
 def plot_metrics(
@@ -157,9 +168,11 @@ def plot_metrics(
         y=estimate,
         color=metric,
         facet_row=metric,
-        markers=n,
+        markers=dict(size=n),
+        hover_data={"n": ':'},
         **kw,
     )
+
     fig.for_each_annotation(lambda a: a.update(text=a.text.split("=")[-1]))
     fig.update_layout(showlegend=False)
 

vetiver/server.py

@@ -6,10 +6,8 @@
 import uvicorn
 import requests
 import pandas as pd
-import numpy as np
 from typing import Callable, Union, List
 
-from . import __version__
 from .vetiver_model import VetiverModel
 from .utils import _jupyter_nb
 
@@ -65,7 +63,7 @@ async def ping():
 
         if self.check_ptype is True:
 
-            @app.post("/predict/")
+            @app.post("/predict")
             async def prediction(
                 input_data: Union[self.model.ptype, List[self.model.ptype]]
             ):
@@ -82,7 +80,7 @@ async def prediction(
 
         elif self.check_ptype is False:
 
-            @app.post("/predict/")
+            @app.post("/predict")
             async def prediction(input_data: Request):
 
                 y = await input_data.json()
@@ -99,17 +97,25 @@ async def rapidoc():
                     <!doctype html>
                     <html>
                         <head>
-                        <meta name="viewport" content="width=device-width,minimum-scale=1,initial-scale=1,user-scalable=yes">
+                        <meta name="viewport"
+                        content="width=device-width,minimum-scale=1,initial-scale=1,user-scalable=yes">
                         <title>RapiDoc</title>
-                        <script type="module" src="https://unpkg.com/rapidoc@9.1.3/dist/rapidoc-min.js"></script>
+                        <script type="module"
+                        src="https://unpkg.com/rapidoc@9.3.3/dist/rapidoc-min.js"></script>
                         </script></head>
                         <body>
                             <rapi-doc spec-url="{self.app.openapi_url[1:]}"
-                            id="thedoc" render-style="read" schema-style="tree"
-                            show-components="true" show-info="true" show-header="true"
+                            id="thedoc"
+                            render-style="read"
+                            schema-style="tree"
+                            show-components="true"
+                            show-info="true"
+                            show-header="true"
                             allow-search="true"
                             show-side-nav="false"
-                            allow-authentication="false" update-route="false" match-type="regex"
+                            allow-authentication="false"
+                            update-route="false"
+                            match-type="regex"
                             theme="light"
                             header-color="#F2C6AC"
                             primary-color = "#8C2D2D">
@@ -143,15 +149,15 @@ def vetiver_post(
         """
         if self.check_ptype is True:
 
-            @self.app.post("/" + endpoint_name + "/")
+            @self.app.post("/" + endpoint_name)
             async def custom_endpoint(input_data: self.model.ptype):
                 y = _prepare_data(input_data)
                 new = endpoint_fx(pd.Series(y))
                 return {endpoint_name: new.tolist()}
 
         else:
 
-            @self.app.post("/" + endpoint_name + "/")
+            @self.app.post("/" + endpoint_name)
             async def custom_endpoint(input_data: Request):
                 y = await input_data.json()
                 new = endpoint_fx(pd.Series(y))
@@ -173,11 +179,13 @@ def run(self, port: int = 8000, host: str = "127.0.0.1", **kw):
         uvicorn.run(self.app, port=port, host=host, **kw)
 
     def _custom_openapi(self):
+        import vetiver
+
         if self.app.openapi_schema:
             return self.app.openapi_schema
         openapi_schema = get_openapi(
             title=self.model.model_name + " model API",
-            version=__version__,
+            version=vetiver.__version__,
             description=self.model.description,
             routes=self.app.routes,
             servers=self.app.servers,
@@ -204,7 +212,7 @@ def predict(endpoint, data: Union[dict, pd.DataFrame, pd.Series], **kw):
     """
     if isinstance(endpoint, testclient.TestClient):
         requester = endpoint
-        endpoint = "/predict/"
+        endpoint = "/predict"
     else:
         requester = requests
 

vetiver/tests/test_add_endpoint.py

@@ -28,7 +28,7 @@ def test_endpoint_adds_ptype():
 
     client = TestClient(app)
     data = {"B": 0, "C": 0, "D": 0}
-    response = client.post("/sum/", json=data)
+    response = client.post("/sum", json=data)
     assert response.status_code == 200, response.text
     assert response.json() == {"sum": 0}, response.json()
 
@@ -38,6 +38,6 @@ def test_endpoint_adds_no_ptype():
 
     client = TestClient(app)
     data = [0, 0, 0]
-    response = client.post("/sum/", json=data)
+    response = client.post("/sum", json=data)
     assert response.status_code == 200, response.text
     assert response.json() == {"sum": 0}, response.json()