# Calibration curve - Logistic Regression
# Create the figure and axes
fig, (ax, hist_ax) = plt.subplots(
    nrows=2, figsize=(8, 8), height_ratios=[2, 1], sharex=True
)

# Plot the calibration curve
display.plot(
    ax=ax,
    hist_ax=hist_ax,
    line_kwargs={"color": "darkblue", "marker": "o", "linewidth": 2},
    ref_line_kwargs={
        "color": "black",
        "linestyle": "--",
        "alpha": 0.8,
        "linewidth": 1.5,
    },
    hist_kwargs={"color": "skyblue", "alpha": 0.5},
)

# Calibration curves for different classifiers
# Compares: Logistic Regression, Naive Bayes, Random Forest, and SVC
# Create plot with custom styling
fig, (ax, hist_ax) = plt.subplots(
    nrows=2, figsize=(10, 10), height_ratios=[2, 1], sharex=True
)

# Plot with custom styling
display.plot(
    ax=ax,
    hist_ax=hist_ax,
    line_kwargs={"linewidth": 2, "alpha": 0.8, "marker": "o", "markersize": 5},
    ref_line_kwargs={
        "color": "black",
        "linestyle": "--",
        "alpha": 0.7,
        "linewidth": 1.5,
    },
    hist_kwargs={"alpha": 0.3},
)

# Calibration curves for uncalibrated vs. calibrated models
# Compares: Random Forest with different calibration methods (uncalibrated, sigmoid, isotonic)
# Create a nice plot
fig, (ax, hist_ax) = plt.subplots(
    nrows=2, figsize=(10, 10), height_ratios=[2, 1], sharex=True
)

# Plot with custom styling
display.plot(
    ax=ax,
    hist_ax=hist_ax,
    line_kwargs={"linewidth": 2, "alpha": 0.8, "marker": "o", "markersize": 6},
    ref_line_kwargs={
        "color": "black",
        "linestyle": "--",
        "alpha": 0.7,
        "linewidth": 1.5,
    },
    hist_kwargs={"alpha": 0.3},
)

# Calibration plots for different strategies and bin counts
# Compares: Different binning strategies (uniform vs quantile) with different bin counts
# Setup the figure with GridSpec
fig = plt.figure(figsize=(12, 10))
gs = GridSpec(3, 2, height_ratios=[2, 1, 1])

# Create a single plot for all calibration curves
ax_calibration_curve = fig.add_subplot(gs[0, :])
ax_calibration_curve.set_title(
    "Calibration plots for different strategies and bin counts"
)

# Calibration plots
# Compares: Multiple classifiers in scikit-learn style layout
# Setup the figure with GridSpec
fig = plt.figure(figsize=(12, 10))
gs = GridSpec(4, 2, height_ratios=[2, 2, 1, 1])

# Create a single plot for all calibration curves
ax_calibration_curve = fig.add_subplot(gs[:2, :])
ax_calibration_curve.set_title("Calibration plots", fontsize=16)

# Calibration curves for multiclass classification
# Shows: Calibration curves for a three-class classification problem
# Setup the figure with GridSpec
fig = plt.figure(figsize=(15, 10))
gs = GridSpec(3, 3, height_ratios=[2, 1, 1])

# Create a single plot for all calibration curves
ax_calibration_curve = fig.add_subplot(gs[0, :])
ax_calibration_curve.set_title(
    "Calibration curves for multiclass classification", fontsize=16
)