[WB-6749] Add yea test demonstrating sklearn fix (#2826)

functional_tests/lightning/train-ddp.yea

@@ -1,6 +1,8 @@
 id: 0.lightning.ddp
 plugin:
   - wandb
+tag:
+  shard: service
 command:
   program: train-ddp.py
 depend:

functional_tests/sklearn/01-plot-calibration-curve-nonbinary.py

@@ -0,0 +1,66 @@
+#!/usr/bin/env python
+"""Demonstrate non-binary plot calibration curve failure
+
+Reproduction for WB-6749.
+
+---
+id: 0.sklearn.01-plot-calibration-curve-nonbinary
+plugin:
+  - wandb
+depend:
+  files:
+    - file: wine.csv
+      source: https://raw.githubusercontent.com/wandb/examples/master/examples/data/wine.csv
+assert:
+  - :wandb:runs_len: 1
+  - :wandb:runs[0][exitcode]: 1
+  - :yea:exit: 1
+  - :op:contains_regex:
+    - :wandb:runs[0][output][stderr]
+    - This function only supports binary classification at the moment and therefore expects labels to be binary
+  - :op:contains:
+    - :wandb:runs[0][telemetry][1]  # imports before
+    - 5  # sklearn
+  - :op:contains:
+    - :wandb:runs[0][telemetry][2]  # imports after
+    - 5  # sklearn
+"""
+
+import numpy as np
+import pandas as pd
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.model_selection import train_test_split
+import wandb
+
+# yea test will grab this
+# data_url = "https://raw.githubusercontent.com/wandb/examples/master/examples/data/wine.csv"
+# !wget {data_url} -O "wine.csv"
+
+# Load data
+wine_quality = pd.read_csv("wine.csv")
+y = wine_quality["quality"]
+y = y.values
+X = wine_quality.drop(["quality"], axis=1)
+X = X.values
+feature_names = wine_quality.columns
+
+X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
+labels = ['one', 'two', 'three', 'four', 'five', 'six', 'seven', 'eight', 'nine', 'ten']
+
+# Train model, get predictions
+model = RandomForestClassifier()
+model.fit(X_train, y_train)
+y_pred = model.predict(X_test)
+y_probas = model.predict_proba(X_test)
+importances = model.feature_importances_
+indices = np.argsort(importances)[::-1]
+
+print(model.n_features_)
+
+run = wandb.init(project='my-scikit-integration')
+
+wandb.sklearn.plot_calibration_curve(model, X_train, y_train, 'RandomForestClassifier')
+
+print(model.n_features_)
+
+outs = model.predict(X_train)