Merge 5a9ba08 into 8c61613

examples/notebooks/constraints.ipynb

@@ -1,23 +1,5 @@
 {
  "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "/Volumes/Workspace/whylogs-python/notebooks\n"
-     ]
-    }
-   ],
-   "source": [
-    "! cd .. && python setup.py develop >/dev/null\n",
-    "! pwd\n"
-   ]
-  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -759,7 +741,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.5"
+   "version": "3.8.8"
   }
  },
  "nbformat": 4,

tests/conftest.py

@@ -50,8 +50,14 @@ def s3_all_config_metadata_path():
 
 
 @pytest.fixture(scope="session")
-def df_lending_club():
+def local_config_path():
+    config_path = os.path.join(_MY_DIR, os.pardir, "testdata", ".whylogs_local.yaml")
+    return config_path
+
 
+@pytest.fixture(scope="session")
+def df_lending_club():
+    # just the top 50
     df = pd.read_csv(os.path.join(_MY_DIR, os.pardir, "testdata", "lending_club_1000.csv"))
     return df.head(50)
 

tests/unit/core/statistics/test_constraints.py

@@ -0,0 +1,85 @@
+import json
+
+import pytest
+
+from whylogs.app.config import load_config
+from whylogs.app.session import session_from_config
+from whylogs.core.statistics.constraints import (
+    DatasetConstraints,
+    Op,
+    SummaryConstraint,
+    ValueConstraint,
+    _summary_funcs1,
+    _value_funcs,
+)
+from whylogs.proto import Op
+from whylogs.util.protobuf import message_to_json
+
+
+def test_value_summary_serialization():
+
+    for each_op, _ in _value_funcs.items():
+
+        value = ValueConstraint(each_op, 3.6)
+        msg_value = value.to_protobuf()
+        json_value = json.loads(message_to_json(msg_value))
+        assert json_value["name"] == "value " + Op.Name(each_op) + " 3.6"
+        assert pytest.approx(json_value["value"], 0.001) == 3.6
+        assert json_value["op"] == Op.Name(each_op)
+        assert json_value["verbose"] == False
+
+    for each_op, _ in _summary_funcs1.items():
+        # constraints may have an optional name
+        sum_constraint = SummaryConstraint("min", each_op, 300000, name="< 30K")
+        msg_sum_const = sum_constraint.to_protobuf()
+        json_summary = json.loads(message_to_json(msg_sum_const))
+
+        assert json_summary["name"] == "< 30K"
+        assert pytest.approx(json_summary["value"], 0.1) == 300000
+        assert json_summary["firstField"] == "min"
+        assert json_summary["op"] == str(Op.Name(each_op))
+        assert json_summary["verbose"] == False
+
+
+def test_value_constraints(df_lending_club, local_config_path):
+
+    conforming_loan = ValueConstraint(Op.LT, 548250)
+    smallest_loan = ValueConstraint(Op.GT, 2500.0, verbose=True)
+
+    high_fico = ValueConstraint(Op.GT, 4000)
+
+    dc = DatasetConstraints(None, value_constraints={"loan_amnt": [conforming_loan, smallest_loan], "fico_range_high": [high_fico]})
+
+    config = load_config(local_config_path)
+    session = session_from_config(config)
+
+    profile = session.log_dataframe(df_lending_club, "test.data", constraints=dc)
+    session.close()
+    report = dc.report()
+
+    assert len(report) == 2
+    print(report)
+    # make sure it checked every value
+    for each_feat in report:
+        for each_constraint in each_feat[1]:
+            assert each_constraint[1] == 50
+
+    assert report[1][1][0][2] == 50
+
+
+def test_summary_constraints(df_lending_club, local_config_path):
+
+    non_negative = SummaryConstraint("min", Op.GE, 0)
+
+    dc = DatasetConstraints(None, summary_constraints={"annual_inc": [non_negative]})
+    config = load_config(local_config_path)
+    session = session_from_config(config)
+    profile = session.log_dataframe(df_lending_club, "test.data", constraints=dc)
+    session.close()
+    report = r = profile.apply_summary_constraints()
+
+    assert len(report) == 1
+    # make sure it checked every value
+    for each_feat in report:
+        for each_constraint in each_feat[1]:
+            assert each_constraint[1] == 1

tests/unit/viz/test_profile_viewer.py

@@ -0,0 +1,24 @@
+import datetime
+import os
+
+import numpy as np
+
+from whylogs.app.config import load_config
+from whylogs.app.session import session_from_config
+from whylogs.viz import profile_viewer
+
+
+def test_profile_viewer(tmpdir, local_config_path):
+
+    config = load_config(local_config_path)
+    session = session_from_config(config)
+
+    with session.logger("mytestytest", dataset_timestamp=datetime.datetime(2021, 6, 2)) as logger:
+        for _ in range(5):
+            logger.log({"uniform_integers": np.random.randint(0, 50)})
+            logger.log({"nulls": None})
+
+        profile = logger.profile
+    result = profile_viewer(profiles=[profile], output_path=tmpdir + "my_test.html")
+    assert os.path.exists(tmpdir + "my_test.html")
+    assert result == tmpdir + "my_test.html"