Skip to content

Latest commit

 

History

History
379 lines (266 loc) · 10 KB

configuration.md

File metadata and controls

379 lines (266 loc) · 10 KB

Configuration

Configuration reference for InsightfulPy v0.2.0.

Table of Contents

Overview

InsightfulPy uses centralized constants to eliminate magic numbers and improve maintainability. All configurable values are defined as module-level constants.

Import Pattern:

from insightfulpy import constants

# Access a constant
threshold = constants.IQR_OUTLIER_MULTIPLIER

Statistical Constants

Quartile and Percentile Values

FIRST_QUARTILE = 0.25
THIRD_QUARTILE = 0.75
QUARTILE_25_PERCENTILE = 25
QUARTILE_75_PERCENTILE = 75

Used for calculating quartiles in statistical summaries and outlier detection.

IQR-Based Outlier Detection

IQR_OUTLIER_MULTIPLIER = 1.5

Standard multiplier for IQR-based outlier detection. Values below Q1 - 1.5*IQR or above Q3 + 1.5*IQR are considered outliers.

Functions using this constant:

  • detect_outliers()
  • iqr_trimmed_mean()
  • interconnected_outliers()
  • comp_num_analysis()

Percentage Conversion

PERCENTAGE_MULTIPLIER = 100

Multiplier for converting decimal values to percentages.

Statistical Precision

DEFAULT_DECIMAL_PLACES = 4
PERCENTAGE_DECIMAL_PLACES = 2
  • DEFAULT_DECIMAL_PLACES: Decimal places for numerical statistics
  • PERCENTAGE_DECIMAL_PLACES: Decimal places for percentage values

Data Quality Constants

Display Limits

DEFAULT_MAX_DISPLAY_OUTLIERS = 10
DEFAULT_HIGH_CARDINALITY_THRESHOLD = 20
  • DEFAULT_MAX_DISPLAY_OUTLIERS: Maximum outlier values shown in output
  • DEFAULT_HIGH_CARDINALITY_THRESHOLD: Threshold for identifying high cardinality categorical columns

Usage:

import insightfulpy as ipy

# Uses DEFAULT_MAX_DISPLAY_OUTLIERS
outliers = ipy.detect_outliers(df)

# Uses DEFAULT_HIGH_CARDINALITY_THRESHOLD
high_card = ipy.cat_high_cardinality(df)

Normality Test Thresholds

MIN_NORMALITY_TEST_SAMPLE_SIZE = 3
SHAPIRO_WILK_MAX_SAMPLE_SIZE = 5000
  • MIN_NORMALITY_TEST_SAMPLE_SIZE: Minimum sample size required for normality testing
  • SHAPIRO_WILK_MAX_SAMPLE_SIZE: Maximum sample size for Shapiro-Wilk test (Kolmogorov-Smirnov test used above this threshold)

Functions affected:

  • comp_num_analysis()

Visualization Constants

Figure Dimensions

DEFAULT_FIGURE_WIDTH = 12
DEFAULT_FIGURE_HEIGHT = 6
LARGE_FIGURE_WIDTH = 18
LARGE_FIGURE_HEIGHT = 8
EXTENDED_FIGURE_WIDTH = 24

Default figure sizes (in inches) for different visualization types:

  • DEFAULT_FIGURE_WIDTH x DEFAULT_FIGURE_HEIGHT: Standard plots
  • LARGE_FIGURE_WIDTH x LARGE_FIGURE_HEIGHT: Larger visualizations
  • EXTENDED_FIGURE_WIDTH: Extra wide plots

Subplot and Layout

DEFAULT_SUBPLOT_COLS = 3
MAX_SUBPLOTS_PER_BATCH = 12
SUBPLOT_HEIGHT_MULTIPLIER = 5
SUBPLOT_WIDTH_MULTIPLIER = 6
  • DEFAULT_SUBPLOT_COLS: Number of columns in subplot grid (3 columns)
  • MAX_SUBPLOTS_PER_BATCH: Maximum subplots per batch (4 rows x 3 columns = 12)
  • SUBPLOT_HEIGHT_MULTIPLIER: Height multiplier for subplot calculation
  • SUBPLOT_WIDTH_MULTIPLIER: Width multiplier for subplot calculation

Batch visualization functions:

  • kde_batches()
  • box_plot_batches()
  • qq_plot_batches()
  • cat_bar_batches()
  • cat_pie_chart_batches()
  • num_vs_num_scatterplot_pair_batch()
  • cat_vs_cat_pair_batch()
  • num_vs_cat_box_violin_pair_batch()

Font and Display Sizes

DEFAULT_FONT_SIZE = 10
LARGE_FONT_SIZE = 12
TITLE_FONT_SIZE = 14
SUPER_TITLE_FONT_SIZE = 20
SMALL_ANNOTATION_SIZE = 8
MEDIUM_ANNOTATION_SIZE = 9

Font sizes for different text elements in visualizations.

Rotation and Positioning

DEFAULT_ROTATION = 45
VERTICAL_ROTATION = 90
ANNOTATION_OFFSET_Y = 3
  • DEFAULT_ROTATION: Default angle for rotated axis labels (45 degrees)
  • VERTICAL_ROTATION: Vertical rotation angle (90 degrees)
  • ANNOTATION_OFFSET_Y: Vertical offset for annotations

Plot Styling

BOX_PLOT_ALPHA = 0.6
VIOLIN_PLOT_ALPHA = 0.3
BOX_PLOT_WIDTH = 0.4
VIOLIN_PLOT_WIDTH = 0.8

Transparency and width settings for box and violin plots.

Functions using these constants:

  • num_vs_cat_box_violin_pair_batch()

Layout Spacing

TIGHT_LAYOUT_PAD = 3.0
SUBPLOT_ADJUST_BOTTOM = 0.3
SUBPLOT_ADJUST_TOP = 0.9
SUBPLOT_WSPACE = 0.5
SUBPLOT_HSPACE = 0.9

Spacing and padding parameters for subplot layouts:

  • TIGHT_LAYOUT_PAD: Padding for tight layout
  • SUBPLOT_ADJUST_BOTTOM: Bottom margin (0.0 to 1.0)
  • SUBPLOT_ADJUST_TOP: Top margin (0.0 to 1.0)
  • SUBPLOT_WSPACE: Width spacing between subplots
  • SUBPLOT_HSPACE: Height spacing between subplots

High Cardinality Limits

Different thresholds for different visualization types:

CAT_VS_CAT_HIGH_CARDINALITY_LIMIT = 19
NUM_VS_CAT_HIGH_CARDINALITY_LIMIT = 20
BAR_CHART_HIGH_CARDINALITY_LIMIT = 19
PIE_CHART_HIGH_CARDINALITY_LIMIT = 20

These limits determine the maximum unique values displayed in categorical visualizations. Categories exceeding these limits are either excluded or grouped as "Other".

Function-specific limits:

  • cat_vs_cat_pair_batch(): Uses CAT_VS_CAT_HIGH_CARDINALITY_LIMIT
  • num_vs_cat_box_violin_pair_batch(): Uses NUM_VS_CAT_HIGH_CARDINALITY_LIMIT
  • cat_bar_batches(): Uses BAR_CHART_HIGH_CARDINALITY_LIMIT
  • cat_pie_chart_batches(): Uses PIE_CHART_HIGH_CARDINALITY_LIMIT

Category Display Limits

MAX_CATEGORIES_FOR_SMALL_LABELS = 5
MAX_CATEGORIES_FOR_DETAILED_DISPLAY = 10
TICK_LABEL_INTERVAL_DIVISOR = 20
  • MAX_CATEGORIES_FOR_SMALL_LABELS: Categories threshold for reducing label font size
  • MAX_CATEGORIES_FOR_DETAILED_DISPLAY: Threshold for detailed category display
  • TICK_LABEL_INTERVAL_DIVISOR: Divisor for calculating tick label intervals

Dynamic Sizing Constants

MIN_DYNAMIC_WIDTH = 12
DYNAMIC_WIDTH_MULTIPLIER = 0.4
BASE_DYNAMIC_HEIGHT = 6
DYNAMIC_HEIGHT_MULTIPLIER = 0.02
FIGURE_WIDTH_MULTIPLIER = 1.5

Constants for calculating dynamic figure sizes based on data characteristics:

  • MIN_DYNAMIC_WIDTH: Minimum width for dynamically sized figures
  • DYNAMIC_WIDTH_MULTIPLIER: Multiplier for width calculation
  • BASE_DYNAMIC_HEIGHT: Base height for dynamic sizing
  • DYNAMIC_HEIGHT_MULTIPLIER: Multiplier for height calculation
  • FIGURE_WIDTH_MULTIPLIER: Additional width multiplier

Functions using dynamic sizing:

  • cat_analyze_and_plot()

Annotation Positioning

ANNOTATION_Y_MULTIPLIER = 0.02
ANNOTATION_X_CENTER_DIVISOR = 2
  • ANNOTATION_Y_MULTIPLIER: Y-axis offset multiplier for annotations
  • ANNOTATION_X_CENTER_DIVISOR: Divisor for centering annotations on X-axis

Grid and Batch Constants

GRID_ROWS_4x3 = 4
GRID_COLS_4x3 = 3
SINGLE_ROW = 1
DOUBLE_COLUMN = 2
ZERO_BASED_INDEX_OFFSET = 1

Grid layout and batch processing constants:

  • GRID_ROWS_4x3: Number of rows in 4x3 grid layout (4 rows)
  • GRID_COLS_4x3: Number of columns in 4x3 grid layout (3 columns)
  • SINGLE_ROW: Single row layout constant (1)
  • DOUBLE_COLUMN: Double column layout constant (2)
  • ZERO_BASED_INDEX_OFFSET: Offset for converting 0-based to 1-based indexing in batch displays (1)

Usage:

The 4x3 grid (4 rows x 3 columns = 12 subplots) determines the MAX_SUBPLOTS_PER_BATCH value used by batch visualization functions. The ZERO_BASED_INDEX_OFFSET is used when displaying batch numbers to users, converting internal 0-based indexing to user-friendly 1-based numbering.

Format Constants

Column Info Formatting

COLUMN_INFO_INDEX_WIDTH = 5
COLUMN_INFO_COL_INDEX_WIDTH = 10
COLUMN_INFO_ATTRIBUTE_WIDTH = 30
COLUMN_INFO_DATA_TYPE_WIDTH = 15
COLUMN_INFO_RANGE_WIDTH = 30
COLUMN_INFO_DISTINCT_WIDTH = 15

Column widths for formatted output in columns_info() function.

Environment Settings

InsightfulPy automatically detects the execution environment and adapts output accordingly. See User Guide - Environment Compatibility for details on supported environments.

Modifying Constants

Accessing Constants

import insightfulpy as ipy
from insightfulpy import constants

# Read current value
print(f"IQR multiplier: {constants.IQR_OUTLIER_MULTIPLIER}")
print(f"Max subplots per batch: {constants.MAX_SUBPLOTS_PER_BATCH}")

Temporary Modification

Constants can be modified at runtime for temporary changes:

from insightfulpy import constants

# Save original value
original_multiplier = constants.IQR_OUTLIER_MULTIPLIER

# Modify for stricter outlier detection
constants.IQR_OUTLIER_MULTIPLIER = 2.0

# Run analysis with modified constant
outliers = ipy.detect_outliers(df)

# Restore original value
constants.IQR_OUTLIER_MULTIPLIER = original_multiplier

Permanent Modification

For permanent changes in development mode:

  1. Install package in editable mode: pip install -e .
  2. Modify constant values at runtime as shown in "Temporary Modification" section
  3. Changes will persist for your local environment

Note: Runtime modifications are session-specific and do not affect the installed package for other users.

Best Practices

  1. Document changes: Add comments explaining why constants were modified
  2. Test thoroughly: Verify behavior after modifying constants
  3. Use context managers: Consider using context managers to restore original values automatically
  4. Scope awareness: Remember that modifications affect all subsequent function calls in the same session

See Also

Version: 0.2.0 | Status: Beta | Python: 3.8-3.12

Copyright 2025 dhaneshbb | License: MIT | Homepage: https://github.com/dhaneshbb/insightfulpy