Skip to content

Sections

Truthound Data Docs reports consist of 9 section types.

Section Types

from truthound.datadocs import SectionType

class SectionType(str, Enum):
    OVERVIEW = "overview"
    COLUMNS = "columns"
    QUALITY = "quality"
    PATTERNS = "patterns"
    DISTRIBUTION = "distribution"
    CORRELATIONS = "correlations"
    RECOMMENDATIONS = "recommendations"
    ALERTS = "alerts"
    CUSTOM = "custom"

Configuring Sections

Default Section Order

from truthound.datadocs import ReportConfig, SectionType

config = ReportConfig(
    sections=[
        SectionType.OVERVIEW,
        SectionType.QUALITY,
        SectionType.COLUMNS,
        SectionType.PATTERNS,
        SectionType.DISTRIBUTION,
        SectionType.CORRELATIONS,
        SectionType.RECOMMENDATIONS,
        SectionType.ALERTS,
    ],
)

Including Specific Sections Only

# Simple report with summary information only
config = ReportConfig(
    sections=[
        SectionType.OVERVIEW,
        SectionType.QUALITY,
        SectionType.ALERTS,
    ],
)

Section Details

1. Overview

Displays key dataset metrics in card format.

Displayed Items: - Row Count: Total number of rows - Column Count: Total number of columns - Memory: Estimated memory usage - Duplicates: Number of duplicate rows (if present) - Missing: Total null cells (if present) - Quality Score: Overall quality score (0-100)

Charts: - Data type distribution (Donut chart)

2. Quality (Data Quality)

Displays scores by quality dimension.

Metrics: - Overall: Overall quality score - Completeness: Data completeness (based on null ratio) - Uniqueness: Uniqueness (based on unique ratio)

Charts: - Missing value distribution by column (Horizontal Bar)

Warnings: - High missing value column alerts - Constant column notifications - Duplicate row warnings

3. Columns (Column Details)

Provides detailed information for each column.

Table: | Column | Type | Null % | Unique % | Distinct | |--------|------|--------|----------|----------| | name | string | 0.0% | 95.2% | 952 | | age | integer | 2.5% | 8.3% | 83 |

Metadata: - Detailed profile data for each column

4. Patterns (Detected Patterns)

Lists automatically detected data patterns.

Displayed Items: - Column: Column where pattern was detected - Pattern: Pattern type (Email, Phone, UUID, etc.) - Regex: Matching regular expression - Match Ratio: Match percentage - Sample Matches: Sample values

This section is hidden if no patterns are detected.

5. Distribution (Value Distribution)

Displays value distribution analysis results.

Charts: - Uniqueness distribution by column (Horizontal Bar)

6. Correlations

Displays correlations between columns.

Display Format: 

column_a ↔ column_b: 0.85 (strong positive correlation)
column_c ↔ column_d: -0.72 (strong negative correlation)

This section is hidden if no correlation data is available.

7. Recommendations

Lists auto-generated improvement suggestions.

Examples: - Add NotNullValidator for column 'email' - Add EmailFormatValidator for column 'email' - Consider implementing duplicate row detection - Review data collection for columns with high missing values

This section is hidden if there are no recommendations.

8. Alerts

Displays data quality issue alerts.

Severity Levels:

Severity Color Description
Info Blue Informational
Warning Yellow Attention needed
Error Red Problem found
Critical Dark Red Serious issue

AlertSpec Structure:

from truthound.datadocs import AlertSpec, SeverityLevel

alert = AlertSpec(
    title="High Missing Values in 'email'",
    message="Column has 65.3% missing values",
    severity=SeverityLevel.WARNING,
    column="email",
    metric="null_ratio",
    value=0.653,
    threshold=0.5,
    suggestion="Consider imputation or removal",
)

This section is hidden if there are no alerts.

9. Custom

Allows inclusion of user-defined content.

SectionSpec

A data class for section configuration.

from truthound.datadocs import SectionSpec, SectionType, ChartSpec, AlertSpec

section = SectionSpec(
    section_type=SectionType.OVERVIEW,
    title="Overview",
    subtitle="Dataset summary and key metrics",
    charts=[ChartSpec(...)],      # Chart list
    tables=[{"headers": [...], "rows": [...]}],  # Table list
    metrics={"row_count": 1000},  # Metrics dictionary
    alerts=[AlertSpec(...)],      # Alert list
    text_blocks=["recommendation 1", "recommendation 2"],  # Text list
    custom_html="",               # Custom HTML
    collapsible=False,            # Whether collapsible
    collapsed_default=False,      # Default collapsed state
    priority=0,                   # Sort priority
    visible=True,                 # Visibility
    metadata={},                  # Additional metadata
)

Section Renderers

Basic Usage

from truthound.datadocs import get_section_renderer, SectionType

# Get section renderer
renderer = get_section_renderer(SectionType.OVERVIEW)

# Render section
html = renderer.render(section_spec, chart_renderer, theme_config)

Available Renderers

from truthound.datadocs import (
    OverviewSection,
    ColumnsSection,
    QualitySection,
    PatternsSection,
    DistributionSection,
    CorrelationsSection,
    RecommendationsSection,
    AlertsSection,
    CustomSection,
)

# Direct usage
renderer = OverviewSection()
html = renderer.render(section_spec, chart_renderer, theme_config)

Custom Section Renderers

Registering a New Section Renderer

from truthound.datadocs import (
    BaseSectionRenderer,
    SectionSpec,
    register_section_renderer,
)

@register_section_renderer("executive_summary")
class ExecutiveSummarySection(BaseSectionRenderer):
    """Executive summary section."""

    section_type = "executive_summary"

    def render(
        self,
        spec: SectionSpec,
        chart_renderer,
        theme_config,
    ) -> str:
        metrics = spec.metrics
        return f"""
        <section class="executive-summary">
            <h2>{spec.title}</h2>
            <div class="key-metrics">
                <div class="metric">
                    <span class="value">{metrics.get('quality_score', 0)}</span>
                    <span class="label">Quality Score</span>
                </div>
                <div class="metric">
                    <span class="value">{len(spec.alerts)}</span>
                    <span class="label">Issues Found</span>
                </div>
            </div>
            <div class="recommendation">
                {spec.text_blocks[0] if spec.text_blocks else ''}
            </div>
        </section>
        """

Using Custom Sections

from truthound.datadocs import ReportConfig, SectionType

config = ReportConfig(
    sections=[
        "executive_summary",  # Custom section (string)
        SectionType.OVERVIEW,
        SectionType.QUALITY,
        SectionType.COLUMNS,
    ]
)

API Reference

SectionType Enum

class SectionType(str, Enum):
    OVERVIEW = "overview"
    COLUMNS = "columns"
    QUALITY = "quality"
    PATTERNS = "patterns"
    DISTRIBUTION = "distribution"
    CORRELATIONS = "correlations"
    RECOMMENDATIONS = "recommendations"
    ALERTS = "alerts"
    CUSTOM = "custom"

SeverityLevel Enum

class SeverityLevel(str, Enum):
    INFO = "info"
    WARNING = "warning"
    ERROR = "error"
    CRITICAL = "critical"

AlertSpec

@dataclass
class AlertSpec:
    title: str                           # Alert title
    message: str                         # Alert message
    severity: SeverityLevel = SeverityLevel.INFO  # Severity
    column: str | None = None            # Related column
    metric: str | None = None            # Related metric name
    value: float | None = None           # Current value
    threshold: float | None = None       # Threshold value
    suggestion: str | None = None        # Improvement suggestion

SectionSpec

@dataclass
class SectionSpec:
    section_type: SectionType            # Section type
    title: str                           # Section title
    subtitle: str = ""                   # Subtitle
    charts: list[ChartSpec]              # Chart list
    tables: list[dict[str, Any]]         # Table list
    metrics: dict[str, Any]              # Metrics dictionary
    alerts: list[AlertSpec]              # Alert list
    text_blocks: list[str]               # Text block list
    custom_html: str = ""                # Custom HTML
    collapsible: bool = False            # Whether collapsible
    collapsed_default: bool = False      # Default collapsed state
    priority: int = 0                    # Sort priority
    visible: bool = True                 # Visibility
    metadata: dict[str, Any]             # Additional metadata

See Also