API Reference

Complete API documentation for clgraph - SQL column lineage analysis and pipeline orchestration.

Quick Start

# Trace lineage - find source columns
sources = pipeline.trace_column_backward("analytics.customer_metrics", "total_amount")
print(f"Source columns: {sources}")

# Forward impact analysis
impacts = pipeline.trace_column_forward("raw.orders", "amount")
print(f"Impacted columns: {impacts}")

# Export to JSON
data = pipeline.to_json()
print(f"Exported {len(data['columns'])} columns")

Core Classes

Pipeline

The main entry point for all lineage operations.

• Pipeline - Create pipelines, trace lineage, manage metadata, execute queries

Lineage

Classes for understanding data flow.

• PipelineLineageGraph - Column-level lineage graph for pipelines
• ColumnNode - Unified column metadata
• ColumnEdge - Lineage relationships
• TableDependencyGraph - Table-level dependencies
• TableNode - Table metadata
• SQLColumnTracer - Single-query column lineage (advanced)

Validation

Classes for SQL quality validation.

• ValidationIssue - Validation issue details
• IssueSeverity - Error, warning, info levels
• IssueCategory - Issue categorization

Export

Export lineage data to various formats.

• JSONExporter - Export to JSON
• CSVExporter - Export to CSV

Visualization

Create GraphViz visualizations.

• Visualization API - Full visualization function reference
• visualize_pipeline_lineage - Multi-query column lineage
• visualize_table_dependencies - Table-level DAG
• visualize_lineage_path - Traced lineage path

Comparison

Track changes between pipeline versions.

• PipelineDiff - Compare pipelines
• ColumnDiff - Individual column changes

LLM-Powered Features

Natural language interfaces and AI-powered tools.

• LineageAgent - Natural language interface to lineage
• Lineage Tools - Building blocks for lineage operations
• GenerateSQLTool - Text-to-SQL generation
• ExplainQueryTool - SQL explanation

Common Imports

from clgraph import (
    # Main entry point
    Pipeline,

    # Single-query lineage
    SQLColumnTracer,

    # Export formats
    JSONExporter,
    CSVExporter,

    # Visualization functions
    visualize_pipeline_lineage,
    visualize_table_dependencies,
    visualize_lineage_path,
)

# LLM features
from clgraph.agent import LineageAgent
from clgraph.tools import (
    TraceBackwardTool,
    TraceForwardTool,
    ListTablesTool,
    GenerateSQLTool,  # Requires LLM
)

SQL Dialects

clgraph supports multiple SQL dialects via sqlglot:

# BigQuery (default)
pipeline = Pipeline.from_sql_list(queries, dialect="bigquery")

# Snowflake
pipeline = Pipeline.from_sql_list(queries, dialect="snowflake")

# PostgreSQL
pipeline = Pipeline.from_sql_list(queries, dialect="postgres")

# Other: mysql, redshift, spark, duckdb, etc.