Add observability and tests

Turn the stable example into repeatable checks. Capture the command, input fixture, expected output, and important boundary. Tests should be fast enough to run before every change and precise enough that a failure identifies behavior rather than just saying the chapter broke.

Save code and dependency versions before considering model serialization. Publish the supported input shape, target units, evaluation set description, baseline comparison, and known extrapolation limits.

from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_absolute_error

X_train = [[1], [2], [3], [4]]
y_train = [52, 60, 68, 76]
X_test, y_test = [[5]], [84]

model = LinearRegression()
model.fit(X_train, y_train)
prediction = model.predict(X_test)
print(f"slope={model.coef_[0]:.1f}")
print(f"prediction={prediction[0]:.1f}")
print(f"mae={mean_absolute_error(y_test, prediction):.1f}")

Expected evidence:

slope=8.0
prediction=84.0
mae=0.0

The output may include version-specific details such as hashes, paths, fitted thresholds, or final decimal places. Compare the structural facts described here rather than copying placeholders. If the structure differs, stop and inspect the earliest unexpected line.

1. scikit-learn expects X as rows by features, so each hour value sits inside a one-item list.
2. fit estimates coefficient and intercept using only training rows.
3. predict applies frozen parameters to the held-out row.
4. mean absolute error measures the average magnitude of mistakes in the target's units.

These lines form one chain: training pairs of study hours and observed scores plus one untouched test pair becomes a fitted slope and intercept, one held-out prediction, and an error value. Change only one input first. When several values change together, you cannot tell which change caused the new behavior.

• First failure: a one-dimensional X_train triggers an expected-2D-array error; reshape or use nested rows. Re-run the smallest command that proves the repair.
• Second failure: calling predict before fit raises a not-fitted error. Preserve the failing input as a test when it represents a realistic mistake.
• Misleading success: mixing the held-out row into fitting makes the evaluation optimistic and defeats the test. A clean-looking final line cannot cancel contradictory intermediate evidence.

When debugging, copy the exact error text and inspect names, paths, shapes, types, and versions. Explain the cause in one sentence before changing code. That discipline prevents a guessed repair from creating a second defect.

On this page, the practical job is to turn important behavior into repeatable checks. The running case is a small linear relationship between study hours and score must be fit with one row held out.

For the current test step, save the smallest useful evidence: the relevant command, its output, and the input that produced it. Do not use a screenshot as the only record when text can be copied and searched. Keep generated artifacts separate from source inputs so rerunning the example does not destroy the evidence it is meant to evaluate.

The deliverable for this step is a fitted linear regression with a held-out prediction, mean absolute error, and recorded coefficient.

Return to your opening prediction. Mark it correct or rewrite it with the condition you missed. Then explain the difference between a successful execution and a trustworthy result for this specific example.