Merge pull request #119 from rstudio/rsc-testing

remove xfail from `test_rsconnect`

Author: isabelizimm

.github/workflows/tests.yml

@@ -51,10 +51,13 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          python -m pip install -e .[dev,torch]
+          python -m pip install ".[dev]"
+          python -m pip install --upgrade git+https://github.com/rstudio/vetiver-python@${{ github.sha }}
+          echo {{ github.sha }}
       - name: run RStudio Connect
         run: |
           docker-compose up --build -d
+          pip freeze > requirements.txt
           make dev
         env:
           RSC_LICENSE: ${{ secrets.RSC_LICENSE }}

docker-compose.yml

@@ -1,9 +1,9 @@
-version: '3.1'
+version: '3.2'
 
 services:
 
   rsconnect:
-    image: rstudio/rstudio-connect:2021.12.1
+    image: rstudio/rstudio-connect:latest
     restart: always
     ports:
       - 3939:3939

script/setup-rsconnect/rstudio-connect.gcfg

@@ -12,7 +12,9 @@ Provider = pam
 DefaultUserRole = publisher
 
 [Python]
-Enabled = false
+Enabled = true
+Executable = /opt/python/3.8.10/bin/python
+Executable = /opt/python/3.9.5/bin/python
 
 [RPackageRepository "CRAN"]
 URL = https://packagemanager.rstudio.com/cran/__linux__/bionic/latest

setup.cfg

@@ -34,6 +34,7 @@ install_requires =
     pins
     rsconnect-python
     plotly
+    pip-tools
     importlib-metadata>=4.4  # NOTE: Remove when python_requires>=3.8
 
 [options.extras_require]

vetiver/attach_pkgs.py

@@ -29,5 +29,5 @@ def load_pkgs(model: VetiverModel = None, packages: list = None, path=""):
         for package in required_pkgs:
             f.write(package + "\n")
 
-    os.system(f"pip-compile {tmp.name} --output-file={path}vetiver_requirements.txt")
+    os.system(f"pip-compile {tmp.name} --output-file={path}requirements.txt")
     os.remove(tmp.name)

vetiver/tests/test_rsconnect.py

@@ -1,27 +1,27 @@
 import pytest
 import json
 import sklearn
+import pins
+import pandas as pd
+import numpy as np
+
 from pins.boards import BoardRsConnect
 from pins.rsconnect.api import RsConnectApi
 from pins.rsconnect.fs import RsConnectFs
-from rsconnect.api import RSConnectServer
+from rsconnect.api import RSConnectServer, RSConnectClient
 
 import vetiver
 
-# Load data, model
-X_df, y = vetiver.mock.get_mock_data()
-model = vetiver.mock.get_mock_model().fit(X_df, y)
-
 RSC_SERVER_URL = "http://localhost:3939"
 RSC_KEYS_FNAME = "vetiver/tests/rsconnect_api_keys.json"
 
 pytestmark = pytest.mark.rsc_test  # noqa
 
 
-def server_from_key(name):
+def get_key(name):
     with open(RSC_KEYS_FNAME) as f:
         api_key = json.load(f)[name]
-        return RSConnectServer(RSC_SERVER_URL, api_key)
+        return api_key
 
 
 def rsc_from_key(name):
@@ -59,27 +59,42 @@ def rsc_short():
     rsc_delete_user_content(fs_susan.api)
 
 
-def test_board_pin_write(rsc_short):
-    v = vetiver.VetiverModel(
-        model=model, ptype_data=X_df, model_name="susan/model", versioned=None
-    )
-    vetiver.vetiver_pin_write(board=rsc_short, model=v)
-    assert isinstance(rsc_short.pin_read("susan/model"), sklearn.dummy.DummyRegressor)
+def test_deploy(rsc_short):
+    np.random.seed(500)
 
+    # Load data, model
+    X_df, y = vetiver.mock.get_mock_data()
+    model = vetiver.mock.get_mock_model().fit(X_df, y)
 
-@pytest.mark.xfail
-def test_deploy(rsc_short):
-    v = vetiver.VetiverModel(
-        model=model, ptype_data=X_df, model_name="susan/model", versioned=None
-    )
+    v = vetiver.VetiverModel(model=model, ptype_data=X_df, model_name="susan/model")
 
-    vetiver.vetiver_pin_write(board=rsc_short, model=v)
-    connect_server = RSConnectServer(
-        url=RSC_SERVER_URL, api_key=server_from_key("susan")
+    board = pins.board_rsconnect(
+        server_url=RSC_SERVER_URL, api_key=get_key("susan"), allow_pickle_read=True
     )
+
+    vetiver.vetiver_pin_write(board=board, model=v)
+    connect_server = RSConnectServer(url=RSC_SERVER_URL, api_key=get_key("susan"))
+    assert isinstance(board.pin_read("susan/model"), sklearn.dummy.DummyRegressor)
+
     vetiver.deploy_rsconnect(
-        connect_server=connect_server, board=rsc_short, pin_name="susan/model"
+        connect_server=connect_server,
+        board=board,
+        pin_name="susan/model",
+        title="testapi",
+        extra_files=["requirements.txt"],
     )
-    response = vetiver.predict(RSC_SERVER_URL + "/predict", json=X_df)
-    assert response.status_code == 200, response.text
-    assert response.json() == {"prediction": [44.47, 44.47]}, response.json()
+
+    # get url of where content lives
+    client = RSConnectClient(connect_server)
+    dicts = client.content_search()
+    rsc_api = list(filter(lambda x: x["title"] == "testapi", dicts))
+    content_url = rsc_api[0].get("content_url")
+
+    h = {"Authorization": f'Key {get_key("susan")}'}
+
+    endpoint = vetiver.vetiver_endpoint(content_url + "/predict")
+    response = vetiver.predict(endpoint, X_df, headers=h)
+
+    assert isinstance(response, pd.DataFrame), response
+    assert response.iloc[0, 0] == 44.47
+    assert len(response) == 100