Merge pull request #65 from isabelizimm/dev-new-names

renaming functions

Author: isabelizimm

vetiver/__init__.py

@@ -6,14 +6,14 @@
 
 import importlib
 from .ptype import *
-from .vetiver_model import *
-from .server import *
-from .mock import *
-from .pin_read_write import *
+from .vetiver_model import VetiverModel
+from .server import VetiverAPI, vetiver_endpoint
+from .mock import get_mock_data, get_mock_model
+from .pin_read_write import vetiver_pin_write
 from .attach_pkgs import *
 from .meta import *
-from .write_docker import *
-from .write_fastapi import * 
+from .write_docker import write_docker
+from .write_fastapi import write_app
 from .handlers._interface import create_handler, InvalidModelError
 from .handlers.base import VetiverHandler
 from .handlers.sklearn import SKLearnHandler

vetiver/handlers/base.py

@@ -1,7 +1,8 @@
-from ..ptype import vetiver_create_ptype
-from ..meta import vetiver_meta
 from abc import ABCMeta
 
+from ..ptype import vetiver_create_ptype
+from ..meta import _model_meta
+
 class VetiverHandler(metaclass=ABCMeta):
     """Base handler class for creating VetiverModel of different type.
 
@@ -30,7 +31,7 @@ def create_meta(
         required_pkgs: list = [],
     ):
         """Create metadata for sklearn model"""
-        meta = vetiver_meta(user, version, url, required_pkgs)
+        meta = _model_meta(user, version, url, required_pkgs)
 
         return meta
 

vetiver/handlers/sklearn.py

@@ -1,8 +1,9 @@
 import pandas as pd
 import sklearn
 
+from ..meta import _model_meta
 from .base import VetiverHandler
-from ..meta import vetiver_meta
+
 
 class SKLearnHandler(VetiverHandler):
     """Handler class for creating VetiverModels with sklearn.
@@ -12,14 +13,14 @@ class SKLearnHandler(VetiverHandler):
     model : sklearn.base.BaseEstimator
         a trained sklearn model
     """
+
     base_class = sklearn.base.BaseEstimator
-    
+
     def __init__(self, model, ptype_data):
         super().__init__(model, ptype_data)
 
     def describe(self):
-        """Create description for sklearn model
-        """
+        """Create description for sklearn model"""
         desc = f"Scikit-learn {self.model.__class__} model"
         return desc
 
@@ -31,11 +32,10 @@ def construct_meta(
     ):
         """Create metadata for sklearn model"""
         required_pkgs = required_pkgs + ["scikit-learn"]
-        meta = vetiver_meta(user, version, url, required_pkgs)
+        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
 
@@ -58,7 +58,7 @@ def handler_predict(self, input_data, check_ptype):
             if isinstance(input_data, pd.DataFrame):
                 prediction = self.model.predict(input_data)
             else:
-               prediction = self.model.predict([input_data])
+                prediction = self.model.predict([input_data])
 
         # do not check ptype
         else:

vetiver/handlers/torch.py

@@ -1,7 +1,7 @@
 import numpy as np
 
+from ..meta import _model_meta
 from .base import VetiverHandler
-from ..meta import vetiver_meta
 
 torch_exists = True
 try:
@@ -18,13 +18,14 @@ class TorchHandler(VetiverHandler):
     model : nn.Module
         a trained torch model
     """
+
     base_class = torch.nn.Module
+
     def __init__(self, model, ptype_data):
         super().__init__(model, ptype_data)
 
     def describe(self):
-        """Create description for torch model
-        """
+        """Create description for torch model"""
         desc = f"Pytorch model of type {type(self.model)}"
         return desc
 
@@ -34,10 +35,9 @@ def create_meta(
         url: str = None,
         required_pkgs: list = [],
     ):
-        """Create metadata for torch model
-        """
+        """Create metadata for torch model"""
         required_pkgs = required_pkgs + ["torch"]
-        meta = vetiver_meta(user, version, url, required_pkgs)
+        meta = _model_meta(user, version, url, required_pkgs)
 
         return meta
 
@@ -62,9 +62,9 @@ def handler_predict(self, input_data, check_ptype):
             if check_ptype == True:
                 input_data = np.array(input_data, dtype=np.array(self.ptype_data).dtype)
                 prediction = self.model(torch.from_numpy(input_data))
-            
+
             # do not check ptype
-            else:    
+            else:
                 input_data = torch.tensor(input_data)
                 prediction = self.model(input_data)
 

vetiver/meta.py

@@ -1,4 +1,4 @@
-def vetiver_meta(user: dict = None, version: str = None,
+def _model_meta(user: dict = None, version: str = None,
             url: str = None, required_pkgs: list = None):
     """Populate relevant metadata for VetiverModel
 

vetiver/pin_read_write.py

@@ -1,16 +1,16 @@
-import pydantic
 import pins
 import warnings
 import json
 
 from .vetiver_model import VetiverModel
-from .meta import vetiver_meta
+from .meta import _model_meta
 from .write_fastapi import _choose_version
 
-def vetiver_pin_write(board, model: VetiverModel, versioned: bool=True):
+
+def vetiver_pin_write(board, model: VetiverModel, versioned: bool = True):
     """
     Pin a trained VetiverModel along with other model metadata.
-    
+
     Parameters
     ----------
     board:
@@ -21,16 +21,18 @@ def vetiver_pin_write(board, model: VetiverModel, versioned: bool=True):
         Whether or not the pin should be versioned
     """
     if not board.allow_pickle_read:
-        raise NotImplementedError # must be pickle-able
+        raise NotImplementedError  # must be pickle-able
 
     board.pin_write(
         model.model,
-        name = model.model_name,
-        type = "joblib",
-        description = model.description,
-        metadata = {"required_pkgs": model.metadata.get("required_pkgs"),
-                    "ptype": None if model.ptype == None else model.ptype().json()},
-        versioned=versioned
+        name=model.model_name,
+        type="joblib",
+        description=model.description,
+        metadata={
+            "required_pkgs": model.metadata.get("required_pkgs"),
+            "ptype": None if model.ptype == None else model.ptype().json(),
+        },
+        versioned=versioned,
     )
 
     # to do: Model card
@@ -46,15 +48,15 @@ def vetiver_pin_write(board, model: VetiverModel, versioned: bool=True):
 def vetiver_pin_read(board, name: str, version: str = None) -> VetiverModel:
     """
     Read pin and populate VetiverModel
-    
+
     Parameters
     ----------
     board:
         A pin board, created by `pins.board_folder()` or another `board_` function.
     name: string
         Pin name
     version: str
-        Retrieve a specific version of a pin. 
+        Retrieve a specific version of a pin.
 
     Returns
     --------
@@ -63,23 +65,22 @@ def vetiver_pin_read(board, name: str, version: str = None) -> VetiverModel:
     Notes
     -----
     If reading a board from RSConnect, the `board` argument must be in "username/modelname" format.
-    
+
     """
-    version = version if version is not None else _choose_version(board.pin_versions(name))
-
-    model = board.pin_read(name, version)
-    meta = board.pin_meta(name)
-
-    v = VetiverModel(model = model,
-        model_name = name,
-        description = meta.description,
-        metadata = vetiver_meta(user = meta.user,
-             version = version,
-             url = meta.user.get("url"), # None all the time, besides Connect
-             required_pkgs = meta.user.get("required_pkgs")
-        ),
-        ptype_data = json.loads(meta.user.get("ptype")) if meta.user.get("ptype") else None,
-        versioned = True
-        )
-    
+
+    warnings.warn(
+        "vetiver_pin_read will be removed in v1.0.0. Use classmethod VetiverModel.from_pin() instead",
+        DeprecationWarning,
+    )
+
+    version = (
+        version if version is not None else _choose_version(board.pin_versions(name))
+    )
+
+    v = VetiverModel.from_pin(
+        board = board, 
+        name = name,
+        version = version
+    )
+
     return v

vetiver/tests/test_build_vetiver_model.py

@@ -1,10 +1,12 @@
 import pytest
+import sklearn
 
-from vetiver.vetiver_model import VetiverModel
+import vetiver as vt
 from vetiver.mock import get_mock_data, get_mock_model
 
 import pandas as pd
-from numpy import int64
+import pydantic
+import pins
 
 # Load data, model
 X_df, y = get_mock_data()
@@ -14,7 +16,7 @@
 
 def test_vetiver_model_array_ptype():
     # build VetiverModel, no ptype
-    vt1 = VetiverModel(
+    vt1 = vt.VetiverModel(
         model=model,
         ptype_data=X_array,
         model_name="iris",
@@ -24,12 +26,13 @@ def test_vetiver_model_array_ptype():
     )
 
     assert vt1.model == model
+    assert isinstance(vt1.ptype.construct(), pydantic.BaseModel)
     assert list(vt1.ptype.__fields__.values())[0].type_ == int
 
 
 def test_vetiver_model_df_ptype():
     # build VetiverModel, df ptype_data
-    vt2 = VetiverModel(
+    vt2 = vt.VetiverModel(
         model=model,
         ptype_data=X_df,
         model_name="iris",
@@ -39,12 +42,13 @@ def test_vetiver_model_df_ptype():
     )
 
     assert vt2.model == model
+    assert isinstance(vt2.ptype.construct(), pydantic.BaseModel)
     assert list(vt2.ptype.__fields__.values())[0].type_ == int
 
 
 def test_vetiver_model_dict_ptype():
     dict_data = {"B": 0, "C": 0, "D": 0}
-    vt3 = VetiverModel(
+    vt3 = vt.VetiverModel(
         model=model,
         ptype_data=dict_data,
         model_name="iris",
@@ -54,12 +58,13 @@ def test_vetiver_model_dict_ptype():
     )
 
     assert vt3.model == model
+    assert isinstance(vt3.ptype.construct(), pydantic.BaseModel)
     assert list(vt3.ptype.__fields__.values())[0].type_ == int
 
 
 def test_vetiver_model_no_ptype():
     # build VetiverModel, no ptype
-    vt4 = VetiverModel(
+    vt4 = vt.VetiverModel(
         model=model,
         ptype_data=None,
         model_name="iris",
@@ -70,3 +75,22 @@ def test_vetiver_model_no_ptype():
 
     assert vt4.model == model
     assert vt4.ptype == None
+
+
+def test_vetiver_model_from_pin():
+
+    v = vt.VetiverModel(
+        model=model,
+        ptype_data=X_df,
+        model_name="model",
+        versioned=None,
+        description=None,
+        metadata=None,
+    )
+    board = pins.board_temp(allow_pickle_read=True)
+    vt.vetiver_pin_write(board=board, model=v)
+    v2 = vt.VetiverModel.from_pin(board, "model")
+    assert isinstance(v2, vt.VetiverModel)
+    assert isinstance(v2.model, sklearn.base.BaseEstimator)
+    assert isinstance(v2.ptype.construct(), pydantic.BaseModel)
+    board.pin_delete("model")