sql-glider 0.1.2__py3-none-any.whl

sqlglider/cli.py

@@ -0,0 +1,1137 @@
+"""CLI entry point for SQL Glider."""
+
+import json
+import sys
+from pathlib import Path
+from typing import Callable, List, Optional
+
+import typer
+from rich.console import Console
+from rich.table import Table
+from sqlglot.errors import ParseError
+from typing_extensions import Annotated
+
+from sqlglider.global_models import AnalysisLevel, NodeFormat
+from sqlglider.lineage.analyzer import LineageAnalyzer
+from sqlglider.lineage.formatters import (
+    CsvFormatter,
+    JsonFormatter,
+    OutputWriter,
+    TableCsvFormatter,
+    TableJsonFormatter,
+    TableTextFormatter,
+    TextFormatter,
+)
+from sqlglider.templating import (
+    TemplaterError,
+    get_templater,
+    list_templaters,
+    load_all_variables,
+)
+from sqlglider.utils.config import load_config
+
+app = typer.Typer(
+    name="sqlglider",
+    help="SQL Utility Toolkit for better understanding, use, and governance of your queries.",
+    invoke_without_command=False,
+)
+console = Console()
+err_console = Console(stderr=True)
+
+
+def _apply_templating(
+    sql: str,
+    templater_name: Optional[str],
+    cli_vars: Optional[List[str]],
+    vars_file: Optional[Path],
+    config,
+    source_path: Optional[Path] = None,
+) -> str:
+    """Apply templating to SQL if a templater is specified.
+
+    Args:
+        sql: The SQL string to template.
+        templater_name: Name of the templater to use (e.g., "jinja").
+                       If None, returns sql unchanged.
+        cli_vars: List of CLI variable strings in "key=value" format.
+        vars_file: Path to a variables file (JSON or YAML).
+        config: The loaded ConfigSettings object.
+        source_path: Path to source file for resolving includes.
+
+    Returns:
+        The templated SQL string, or the original if no templater specified.
+    """
+    if not templater_name:
+        return sql
+
+    # Get variables from config
+    config_vars_file = None
+    config_vars = None
+    if config.templating:
+        if config.templating.variables_file and not vars_file:
+            config_vars_file = Path(config.templating.variables_file)
+            if not config_vars_file.exists():
+                err_console.print(
+                    f"[yellow]Warning:[/yellow] Variables file from config "
+                    f"not found: {config_vars_file}"
+                )
+                config_vars_file = None
+        config_vars = config.templating.variables
+
+    # Load variables from all sources
+    variables = load_all_variables(
+        cli_vars=cli_vars,
+        vars_file=vars_file or config_vars_file,
+        config_vars=config_vars,
+        use_env=True,
+    )
+
+    # Get templater instance and render
+    templater_instance = get_templater(templater_name)
+    return templater_instance.render(sql, variables=variables, source_path=source_path)
+
+
+@app.callback()
+def main():
+    """SQL Glider - SQL Utility Toolkit."""
+    pass
+
+
+@app.command()
+def lineage(
+    sql_file: Annotated[
+        typer.FileText,
+        typer.Argument(
+            default_factory=lambda: sys.stdin,
+            show_default="stdin",
+            help="Path to SQL file to analyze (reads from stdin if not provided)",
+        ),
+    ],
+    level: Optional[str] = typer.Option(
+        None,
+        "--level",
+        "-l",
+        help="Analysis level: 'column' or 'table' (default: column, or from config)",
+    ),
+    dialect: Optional[str] = typer.Option(
+        None,
+        "--dialect",
+        "-d",
+        help="SQL dialect (default: spark, or from config)",
+    ),
+    column: Optional[str] = typer.Option(
+        None,
+        "--column",
+        "-c",
+        help="Specific output column for forward lineage (default: all columns)",
+    ),
+    source_column: Optional[str] = typer.Option(
+        None,
+        "--source-column",
+        "-s",
+        help="Source column for reverse lineage (impact analysis)",
+    ),
+    table_filter: Optional[str] = typer.Option(
+        None,
+        "--table",
+        help="Filter to only queries that reference this table (for multi-query files)",
+    ),
+    output_format: Optional[str] = typer.Option(
+        None,
+        "--output-format",
+        "-f",
+        help="Output format: 'text', 'json', or 'csv' (default: text, or from config)",
+    ),
+    output_file: Optional[Path] = typer.Option(
+        None,
+        "--output-file",
+        "-o",
+        help="Write output to file instead of stdout",
+    ),
+    templater: Optional[str] = typer.Option(
+        None,
+        "--templater",
+        "-t",
+        help="Templater for SQL preprocessing (e.g., 'jinja', 'none')",
+    ),
+    var: Optional[List[str]] = typer.Option(
+        None,
+        "--var",
+        "-v",
+        help="Template variable in key=value format (repeatable)",
+    ),
+    vars_file: Optional[Path] = typer.Option(
+        None,
+        "--vars-file",
+        exists=True,
+        help="Path to variables file (JSON or YAML)",
+    ),
+) -> None:
+    """
+    Analyze column or table lineage for a SQL file.
+
+    Configuration can be set in sqlglider.toml in the current directory.
+    CLI arguments override configuration file values.
+
+    Examples:
+
+        # Forward lineage: Find sources for output column
+        sqlglider lineage query.sql --column order_total
+
+        # Reverse lineage: Find outputs affected by source column
+        sqlglider lineage query.sql --source-column orders.customer_id
+
+        # Analyze all columns (forward lineage)
+        sqlglider lineage query.sql
+
+        # Analyze table-level lineage
+        sqlglider lineage query.sql --level table
+
+        # Export to JSON
+        sqlglider lineage query.sql --output-format json --output-file lineage.json
+
+        # Use different SQL dialect
+        sqlglider lineage query.sql --dialect postgres
+
+        # Analyze templated SQL with Jinja2
+        sqlglider lineage query.sql --templater jinja --var schema=analytics
+
+        # Use variables file for templating
+        sqlglider lineage query.sql --templater jinja --vars-file vars.json
+    """
+    # Load configuration from sqlglider.toml (if it exists)
+    config = load_config()
+
+    # Apply priority resolution: CLI args > config > defaults
+    dialect = dialect or config.dialect or "spark"
+    level_str = level or config.level or "column"
+    output_format = output_format or config.output_format or "text"
+    templater = templater or config.templater  # None means no templating
+    # Validate and convert level to enum
+    try:
+        analysis_level = AnalysisLevel(level_str)
+    except ValueError:
+        err_console.print(
+            f"[red]Error:[/red] Invalid level '{level_str}'. Use 'column' or 'table'."
+        )
+        raise typer.Exit(1)
+
+    # Validate output format
+    if output_format not in ["text", "json", "csv"]:
+        err_console.print(
+            f"[red]Error:[/red] Invalid output format '{output_format}'. "
+            "Use 'text', 'json', or 'csv'."
+        )
+        raise typer.Exit(1)
+
+    # Validate mutual exclusivity of column and source_column
+    if column and source_column:
+        err_console.print(
+            "[red]Error:[/red] Cannot specify both --column and --source-column. "
+            "Use --column for forward lineage or --source-column for reverse lineage."
+        )
+        raise typer.Exit(1)
+
+    # Check if reading from stdin (cross-platform: name is "<stdin>" on all OS)
+    is_stdin = sql_file.name == "<stdin>"
+
+    try:
+        # Check if stdin is being used without input
+        if is_stdin and sys.stdin.isatty():
+            err_console.print(
+                "[red]Error:[/red] No SQL file provided and stdin is interactive. "
+                "Provide a SQL file path or pipe SQL via stdin."
+            )
+            raise typer.Exit(1)
+
+        # Read SQL from file or stdin
+        sql = sql_file.read()
+
+        # Determine source path for templating (None if stdin)
+        source_path = None if is_stdin else Path(sql_file.name)
+
+        # Apply templating if specified
+        sql = _apply_templating(
+            sql,
+            templater_name=templater,
+            cli_vars=var,
+            vars_file=vars_file,
+            config=config,
+            source_path=source_path,
+        )
+
+        # Create analyzer
+        analyzer = LineageAnalyzer(sql, dialect=dialect)
+
+        # Unified lineage analysis (handles both single and multi-query files)
+        results = analyzer.analyze_queries(
+            level=analysis_level,
+            column=column,
+            source_column=source_column,
+            table_filter=table_filter,
+        )
+
+        # Print warnings for skipped queries
+        for skipped in analyzer.skipped_queries:
+            err_console.print(
+                f"[yellow]Warning:[/yellow] Skipping query {skipped.query_index} "
+                f"({skipped.statement_type}): {skipped.reason}"
+            )
+
+        # Format and output based on output format
+        if output_format == "text":
+            if output_file:
+                # For file output, use a string-based console to capture output
+                from io import StringIO
+
+                from rich.console import Console as FileConsole
+
+                string_buffer = StringIO()
+                file_console = FileConsole(file=string_buffer, force_terminal=False)
+                TextFormatter.format(results, file_console)
+                output_file.write_text(string_buffer.getvalue(), encoding="utf-8")
+                console.print(
+                    f"[green]Success:[/green] Lineage written to {output_file}"
+                )
+            else:
+                # Direct console output with Rich formatting
+                TextFormatter.format(results, console)
+        elif output_format == "json":
+            formatted = JsonFormatter.format(results)
+            OutputWriter.write(formatted, output_file)
+            if output_file:
+                console.print(
+                    f"[green]Success:[/green] Lineage written to {output_file}"
+                )
+        else:  # csv
+            formatted = CsvFormatter.format(results)
+            OutputWriter.write(formatted, output_file)
+            if output_file:
+                console.print(
+                    f"[green]Success:[/green] Lineage written to {output_file}"
+                )
+
+    except FileNotFoundError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ParseError as e:
+        err_console.print(f"[red]Error:[/red] Failed to parse SQL: {e}")
+        raise typer.Exit(1)
+
+    except TemplaterError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ValueError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except Exception as e:
+        err_console.print(f"[red]Error:[/red] Unexpected error: {e}")
+        raise typer.Exit(1)
+
+
+@app.command()
+def tables(
+    sql_file: Annotated[
+        typer.FileText,
+        typer.Argument(
+            default_factory=lambda: sys.stdin,
+            show_default="stdin",
+            help="Path to SQL file to analyze (reads from stdin if not provided)",
+        ),
+    ],
+    dialect: Optional[str] = typer.Option(
+        None,
+        "--dialect",
+        "-d",
+        help="SQL dialect (default: spark, or from config)",
+    ),
+    table_filter: Optional[str] = typer.Option(
+        None,
+        "--table",
+        help="Filter to only queries that reference this table (for multi-query files)",
+    ),
+    output_format: Optional[str] = typer.Option(
+        None,
+        "--output-format",
+        "-f",
+        help="Output format: 'text', 'json', or 'csv' (default: text, or from config)",
+    ),
+    output_file: Optional[Path] = typer.Option(
+        None,
+        "--output-file",
+        "-o",
+        help="Write output to file instead of stdout",
+    ),
+    templater: Optional[str] = typer.Option(
+        None,
+        "--templater",
+        "-t",
+        help="Templater for SQL preprocessing (e.g., 'jinja', 'none')",
+    ),
+    var: Optional[List[str]] = typer.Option(
+        None,
+        "--var",
+        "-v",
+        help="Template variable in key=value format (repeatable)",
+    ),
+    vars_file: Optional[Path] = typer.Option(
+        None,
+        "--vars-file",
+        exists=True,
+        help="Path to variables file (JSON or YAML)",
+    ),
+) -> None:
+    """
+    List all tables involved in a SQL file.
+
+    Outputs table names with their usage type (INPUT, OUTPUT, or BOTH) and
+    object type (TABLE, VIEW, CTE, or UNKNOWN).
+
+    Configuration can be set in sqlglider.toml in the current directory.
+    CLI arguments override configuration file values.
+
+    Examples:
+
+        # List all tables in a SQL file
+        sqlglider tables query.sql
+
+        # Export to JSON
+        sqlglider tables query.sql --output-format json
+
+        # Export to CSV file
+        sqlglider tables query.sql --output-format csv --output-file tables.csv
+
+        # Use different SQL dialect
+        sqlglider tables query.sql --dialect postgres
+
+        # Filter to queries referencing a specific table
+        sqlglider tables query.sql --table customers
+
+        # Analyze templated SQL with Jinja2
+        sqlglider tables query.sql --templater jinja --var schema=analytics
+    """
+    # Load configuration from sqlglider.toml (if it exists)
+    config = load_config()
+
+    # Apply priority resolution: CLI args > config > defaults
+    dialect = dialect or config.dialect or "spark"
+    output_format = output_format or config.output_format or "text"
+    templater = templater or config.templater  # None means no templating
+
+    # Validate output format
+    if output_format not in ["text", "json", "csv"]:
+        err_console.print(
+            f"[red]Error:[/red] Invalid output format '{output_format}'. "
+            "Use 'text', 'json', or 'csv'."
+        )
+        raise typer.Exit(1)
+
+    # Check if reading from stdin (cross-platform: name is "<stdin>" on all OS)
+    is_stdin = sql_file.name == "<stdin>"
+
+    try:
+        # Check if stdin is being used without input
+        if is_stdin and sys.stdin.isatty():
+            err_console.print(
+                "[red]Error:[/red] No SQL file provided and stdin is interactive. "
+                "Provide a SQL file path or pipe SQL via stdin."
+            )
+            raise typer.Exit(1)
+
+        # Read SQL from file or stdin
+        sql = sql_file.read()
+
+        # Determine source path for templating (None if stdin)
+        source_path = None if is_stdin else Path(sql_file.name)
+
+        # Apply templating if specified
+        sql = _apply_templating(
+            sql,
+            templater_name=templater,
+            cli_vars=var,
+            vars_file=vars_file,
+            config=config,
+            source_path=source_path,
+        )
+
+        # Create analyzer
+        analyzer = LineageAnalyzer(sql, dialect=dialect)
+
+        # Analyze tables
+        results = analyzer.analyze_tables(table_filter=table_filter)
+
+        # Format and output based on output format
+        if output_format == "text":
+            if output_file:
+                # For file output, use a string-based console to capture output
+                from io import StringIO
+
+                from rich.console import Console as FileConsole
+
+                string_buffer = StringIO()
+                file_console = FileConsole(file=string_buffer, force_terminal=False)
+                TableTextFormatter.format(results, file_console)
+                output_file.write_text(string_buffer.getvalue(), encoding="utf-8")
+                console.print(
+                    f"[green]Success:[/green] Tables written to {output_file}"
+                )
+            else:
+                # Direct console output with Rich formatting
+                TableTextFormatter.format(results, console)
+        elif output_format == "json":
+            formatted = TableJsonFormatter.format(results)
+            OutputWriter.write(formatted, output_file)
+            if output_file:
+                console.print(
+                    f"[green]Success:[/green] Tables written to {output_file}"
+                )
+        else:  # csv
+            formatted = TableCsvFormatter.format(results)
+            OutputWriter.write(formatted, output_file)
+            if output_file:
+                console.print(
+                    f"[green]Success:[/green] Tables written to {output_file}"
+                )
+
+    except FileNotFoundError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ParseError as e:
+        err_console.print(f"[red]Error:[/red] Failed to parse SQL: {e}")
+        raise typer.Exit(1)
+
+    except TemplaterError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ValueError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except Exception as e:
+        err_console.print(f"[red]Error:[/red] Unexpected error: {e}")
+        raise typer.Exit(1)
+
+
+@app.command()
+def template(
+    sql_file: Annotated[
+        typer.FileText,
+        typer.Argument(
+            default_factory=lambda: sys.stdin,
+            show_default="stdin",
+            help="Path to SQL template file to render (reads from stdin if not provided)",
+        ),
+    ],
+    templater: Optional[str] = typer.Option(
+        None,
+        "--templater",
+        "-t",
+        help="Templater to use (default: jinja, or from config)",
+    ),
+    var: Optional[List[str]] = typer.Option(
+        None,
+        "--var",
+        "-v",
+        help="Template variable in key=value format (repeatable)",
+    ),
+    vars_file: Optional[Path] = typer.Option(
+        None,
+        "--vars-file",
+        exists=True,
+        help="Path to variables file (JSON or YAML)",
+    ),
+    output_file: Optional[Path] = typer.Option(
+        None,
+        "--output-file",
+        "-o",
+        help="Write output to file instead of stdout",
+    ),
+    list_available: bool = typer.Option(
+        False,
+        "--list",
+        "-l",
+        help="List available templaters and exit",
+    ),
+) -> None:
+    """
+    Render a SQL template file with variable substitution.
+
+    Uses the specified templater (default: jinja) to process the SQL file
+    with template variables. Variables can be provided via CLI, file, or config.
+
+    Configuration can be set in sqlglider.toml in the current directory.
+    CLI arguments override configuration file values.
+
+    Examples:
+
+        # Basic template rendering
+        sqlglider template query.sql --var schema=analytics --var table=users
+
+        # Using a variables file
+        sqlglider template query.sql --vars-file vars.json
+
+        # Output to file
+        sqlglider template query.sql --var schema=prod -o rendered.sql
+
+        # List available templaters
+        sqlglider template query.sql --list
+
+        # Use specific templater
+        sqlglider template query.sql --templater jinja --var name=test
+    """
+    # Handle --list option
+    if list_available:
+        available = list_templaters()
+        if available:
+            console.print("[bold]Available templaters:[/bold]")
+            for name in available:
+                console.print(f"  - {name}")
+        else:
+            console.print("[yellow]No templaters available[/yellow]")
+        raise typer.Exit(0)
+
+    # Load configuration from sqlglider.toml (if it exists)
+    config = load_config()
+
+    # Apply priority resolution: CLI args > config > defaults
+    # For template command, default to "jinja" (always apply templating)
+    templater = templater or config.templater or "jinja"
+
+    # Check if reading from stdin (cross-platform: name is "<stdin>" on all OS)
+    is_stdin = sql_file.name == "<stdin>"
+
+    try:
+        # Check if stdin is being used without input
+        if is_stdin and sys.stdin.isatty():
+            err_console.print(
+                "[red]Error:[/red] No SQL file provided and stdin is interactive. "
+                "Provide a SQL file path or pipe SQL via stdin."
+            )
+            raise typer.Exit(1)
+
+        # Read SQL from file or stdin
+        sql = sql_file.read()
+
+        # Determine source path for templating (None if stdin)
+        source_path = None if is_stdin else Path(sql_file.name)
+
+        # Apply templating (always for template command)
+        rendered = _apply_templating(
+            sql,
+            templater_name=templater,
+            cli_vars=var,
+            vars_file=vars_file,
+            config=config,
+            source_path=source_path,
+        )
+
+        # Write output
+        if output_file:
+            output_file.write_text(rendered, encoding="utf-8")
+            console.print(
+                f"[green]Success:[/green] Rendered SQL written to {output_file}"
+            )
+        else:
+            print(rendered)
+
+    except FileNotFoundError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except TemplaterError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ValueError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except Exception as e:
+        err_console.print(f"[red]Error:[/red] Unexpected error: {e}")
+        raise typer.Exit(1)
+
+
+# Graph command group
+graph_app = typer.Typer(
+    name="graph",
+    help="Graph-based lineage analysis commands.",
+)
+app.add_typer(graph_app, name="graph")
+
+
+@graph_app.command("build")
+def graph_build(
+    paths: List[Path] = typer.Argument(
+        None,
+        help="SQL file(s) or directory path to process",
+    ),
+    output: Path = typer.Option(
+        ...,
+        "--output",
+        "-o",
+        help="Output file path for serialized graph (required)",
+    ),
+    recursive: bool = typer.Option(
+        False,
+        "--recursive",
+        "-r",
+        help="Recursively search directories for SQL files",
+    ),
+    glob_pattern: str = typer.Option(
+        "*.sql",
+        "--glob",
+        "-g",
+        help="Glob pattern for matching SQL files in directories",
+    ),
+    manifest: Optional[Path] = typer.Option(
+        None,
+        "--manifest",
+        "-m",
+        exists=True,
+        help="Path to manifest CSV file with file_path and optional dialect columns",
+    ),
+    node_format: str = typer.Option(
+        "qualified",
+        "--node-format",
+        "-n",
+        help="Node identifier format: 'qualified' or 'structured'",
+    ),
+    dialect: Optional[str] = typer.Option(
+        None,
+        "--dialect",
+        "-d",
+        help="SQL dialect (default: spark, falls back if not in manifest)",
+    ),
+    templater: Optional[str] = typer.Option(
+        None,
+        "--templater",
+        "-t",
+        help="Templater for SQL preprocessing (e.g., 'jinja', 'none')",
+    ),
+    var: Optional[List[str]] = typer.Option(
+        None,
+        "--var",
+        "-v",
+        help="Template variable in key=value format (repeatable)",
+    ),
+    vars_file: Optional[Path] = typer.Option(
+        None,
+        "--vars-file",
+        exists=True,
+        help="Path to variables file (JSON or YAML)",
+    ),
+) -> None:
+    """
+    Build a lineage graph from SQL files.
+
+    Supports multiple input modes:
+    - Single file: sqlglider graph build query.sql -o graph.json
+    - Multiple files: sqlglider graph build query1.sql query2.sql -o graph.json
+    - Directory: sqlglider graph build ./queries/ -r -o graph.json
+    - Manifest: sqlglider graph build --manifest manifest.csv -o graph.json
+
+    Examples:
+
+        # Build from single file
+        sqlglider graph build query.sql -o graph.json
+
+        # Build from directory (recursive)
+        sqlglider graph build ./queries/ -r -o graph.json
+
+        # Build from manifest with custom dialect
+        sqlglider graph build --manifest manifest.csv -o graph.json --dialect postgres
+
+        # Build with structured node format
+        sqlglider graph build query.sql -o graph.json --node-format structured
+
+        # Build with Jinja2 templating
+        sqlglider graph build ./queries/ -o graph.json --templater jinja --var schema=prod
+    """
+    from sqlglider.graph.builder import GraphBuilder
+    from sqlglider.graph.serialization import save_graph
+
+    # Load config for defaults
+    config = load_config()
+    dialect = dialect or config.dialect or "spark"
+    templater = templater or config.templater  # None means no templating
+
+    # Validate and convert node format to enum
+    try:
+        node_format_enum = NodeFormat(node_format)
+    except ValueError:
+        err_console.print(
+            f"[red]Error:[/red] Invalid node format '{node_format}'. "
+            "Use 'qualified' or 'structured'."
+        )
+        raise typer.Exit(1)
+
+    # Validate inputs
+    if not paths and not manifest:
+        err_console.print(
+            "[red]Error:[/red] Must provide either file/directory paths or --manifest option."
+        )
+        raise typer.Exit(1)
+
+    # Create SQL preprocessor if templating is enabled
+    sql_preprocessor: Optional[Callable[[str, Path], str]] = None
+    if templater:
+        # Load variables once for all files
+        config_vars_file = None
+        config_vars = None
+        if config.templating:
+            if config.templating.variables_file and not vars_file:
+                config_vars_file = Path(config.templating.variables_file)
+                if not config_vars_file.exists():
+                    err_console.print(
+                        f"[yellow]Warning:[/yellow] Variables file from config "
+                        f"not found: {config_vars_file}"
+                    )
+                    config_vars_file = None
+            config_vars = config.templating.variables
+
+        variables = load_all_variables(
+            cli_vars=var,
+            vars_file=vars_file or config_vars_file,
+            config_vars=config_vars,
+            use_env=True,
+        )
+
+        templater_instance = get_templater(templater)
+
+        def _preprocess(sql: str, file_path: Path) -> str:
+            return templater_instance.render(
+                sql, variables=variables, source_path=file_path
+            )
+
+        sql_preprocessor = _preprocess
+
+    try:
+        builder = GraphBuilder(
+            node_format=node_format_enum,
+            dialect=dialect,
+            sql_preprocessor=sql_preprocessor,
+        )
+
+        # Process manifest if provided
+        if manifest:
+            builder.add_manifest(manifest, dialect=dialect)
+
+        # Process paths
+        if paths:
+            for path in paths:
+                if path.is_dir():
+                    builder.add_directory(
+                        path,
+                        recursive=recursive,
+                        glob_pattern=glob_pattern,
+                        dialect=dialect,
+                    )
+                elif path.is_file():
+                    builder.add_file(path, dialect=dialect)
+                else:
+                    err_console.print(f"[red]Error:[/red] Path not found: {path}")
+                    raise typer.Exit(1)
+
+        # Build and save graph
+        graph = builder.build()
+        save_graph(graph, output)
+
+        console.print(
+            f"[green]Success:[/green] Graph saved to {output} "
+            f"({graph.metadata.total_nodes} nodes, {graph.metadata.total_edges} edges)"
+        )
+
+    except FileNotFoundError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ParseError as e:
+        err_console.print(f"[red]Error:[/red] Failed to parse SQL: {e}")
+        raise typer.Exit(1)
+
+    except TemplaterError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ValueError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except Exception as e:
+        err_console.print(f"[red]Error:[/red] Unexpected error: {e}")
+        raise typer.Exit(1)
+
+
+@graph_app.command("merge")
+def graph_merge(
+    inputs: List[Path] = typer.Argument(
+        None,
+        help="JSON graph files to merge",
+    ),
+    output: Path = typer.Option(
+        ...,
+        "--output",
+        "-o",
+        help="Output file path for merged graph (required)",
+    ),
+    glob_pattern: Optional[str] = typer.Option(
+        None,
+        "--glob",
+        "-g",
+        help="Glob pattern for matching graph JSON files (e.g., 'graphs/*.json')",
+    ),
+) -> None:
+    """
+    Merge multiple lineage graphs into one.
+
+    Nodes are deduplicated by identifier. Edges are deduplicated by source-target pair.
+
+    Examples:
+
+        # Merge specific files
+        sqlglider graph merge graph1.json graph2.json -o merged.json
+
+        # Merge with glob pattern
+        sqlglider graph merge --glob "graphs/*.json" -o merged.json
+
+        # Combine both
+        sqlglider graph merge extra.json --glob "graphs/*.json" -o merged.json
+    """
+    from sqlglider.graph.merge import GraphMerger
+    from sqlglider.graph.serialization import save_graph
+
+    # Validate inputs
+    if not inputs and not glob_pattern:
+        err_console.print(
+            "[red]Error:[/red] Must provide either graph files or --glob option."
+        )
+        raise typer.Exit(1)
+
+    try:
+        merger = GraphMerger()
+
+        # Process glob pattern if provided
+        if glob_pattern:
+            glob_files = sorted(Path(".").glob(glob_pattern))
+            if not glob_files:
+                err_console.print(
+                    f"[yellow]Warning:[/yellow] No files matched pattern: {glob_pattern}"
+                )
+            for graph_file in glob_files:
+                if graph_file.is_file():
+                    merger.add_file(graph_file)
+
+        # Process explicit inputs
+        if inputs:
+            for graph_file in inputs:
+                if not graph_file.exists():
+                    err_console.print(f"[red]Error:[/red] File not found: {graph_file}")
+                    raise typer.Exit(1)
+                merger.add_file(graph_file)
+
+        # Merge and save
+        merged_graph = merger.merge()
+        save_graph(merged_graph, output)
+
+        console.print(
+            f"[green]Success:[/green] Merged graph saved to {output} "
+            f"({merged_graph.metadata.total_nodes} nodes, {merged_graph.metadata.total_edges} edges, "
+            f"{len(merged_graph.metadata.source_files)} source files)"
+        )
+
+    except FileNotFoundError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ValueError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except Exception as e:
+        err_console.print(f"[red]Error:[/red] Unexpected error: {e}")
+        raise typer.Exit(1)
+
+
+@graph_app.command("query")
+def graph_query(
+    graph_file: Path = typer.Argument(
+        ...,
+        exists=True,
+        help="Path to graph JSON file",
+    ),
+    upstream: Optional[str] = typer.Option(
+        None,
+        "--upstream",
+        "-u",
+        help="Find all source columns that contribute to this column",
+    ),
+    downstream: Optional[str] = typer.Option(
+        None,
+        "--downstream",
+        "-d",
+        help="Find all columns affected by this source column",
+    ),
+    output_format: str = typer.Option(
+        "text",
+        "--output-format",
+        "-f",
+        help="Output format: 'text', 'json', or 'csv'",
+    ),
+) -> None:
+    """
+    Query a lineage graph for upstream or downstream dependencies.
+
+    Examples:
+
+        # Find all source columns for a target
+        sqlglider graph query graph.json --upstream orders.customer_id
+
+        # Find all columns affected by a source
+        sqlglider graph query graph.json --downstream customers.customer_id
+
+        # JSON output
+        sqlglider graph query graph.json --upstream orders.total -f json
+
+        # CSV output
+        sqlglider graph query graph.json --downstream orders.order_id -f csv
+    """
+    from sqlglider.graph.query import GraphQuerier
+
+    # Validate options
+    if not upstream and not downstream:
+        err_console.print(
+            "[red]Error:[/red] Must specify either --upstream or --downstream."
+        )
+        raise typer.Exit(1)
+
+    if upstream and downstream:
+        err_console.print(
+            "[red]Error:[/red] Cannot specify both --upstream and --downstream. "
+            "Choose one direction."
+        )
+        raise typer.Exit(1)
+
+    if output_format not in ["text", "json", "csv"]:
+        err_console.print(
+            f"[red]Error:[/red] Invalid output format '{output_format}'. "
+            "Use 'text', 'json', or 'csv'."
+        )
+        raise typer.Exit(1)
+
+    try:
+        querier = GraphQuerier.from_file(graph_file)
+
+        if upstream:
+            result = querier.find_upstream(upstream)
+        else:
+            assert downstream is not None  # Validated above
+            result = querier.find_downstream(downstream)
+
+        # Format and output
+        if output_format == "text":
+            _format_query_result_text(result)
+        elif output_format == "json":
+            _format_query_result_json(result)
+        else:  # csv
+            _format_query_result_csv(result)
+
+    except FileNotFoundError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except ValueError as e:
+        err_console.print(f"[red]Error:[/red] {e}")
+        raise typer.Exit(1)
+
+    except Exception as e:
+        err_console.print(f"[red]Error:[/red] Unexpected error: {e}")
+        raise typer.Exit(1)
+
+
+def _format_query_result_text(result) -> None:
+    """Format query result as text table."""
+    direction_label = (
+        "Sources" if result.direction == "upstream" else "Affected Columns"
+    )
+
+    table = Table(title=f"{direction_label} for '{result.query_column}'")
+    table.add_column("Column", style="cyan")
+    table.add_column("Table", style="green")
+    table.add_column("Hops", style="yellow", justify="right")
+    table.add_column("Root", style="magenta", justify="center")
+    table.add_column("Leaf", style="magenta", justify="center")
+    table.add_column("Paths", style="dim")
+    table.add_column("File", style="dim")
+
+    for node in result.related_columns:
+        # Format paths as newline-separated arrow strings
+        paths_str = (
+            "\n".join(p.to_arrow_string() for p in node.paths) if node.paths else "-"
+        )
+
+        table.add_row(
+            node.column or node.identifier,
+            node.table or "",
+            str(node.hops),
+            "Y" if node.is_root else "N",
+            "Y" if node.is_leaf else "N",
+            paths_str,
+            Path(node.file_path).name if node.file_path else "",
+        )
+
+    if len(result) == 0:
+        console.print(
+            f"[yellow]No {direction_label.lower()} found for '{result.query_column}'[/yellow]"
+        )
+    else:
+        console.print(table)
+        console.print(f"\n[dim]Total: {len(result)} column(s)[/dim]")
+
+
+def _format_query_result_json(result) -> None:
+    """Format query result as JSON."""
+    columns = []
+    for node in result.related_columns:
+        node_data = node.model_dump()
+        # Serialize paths as arrays of node identifiers for cleaner output
+        node_data["paths"] = [p.nodes for p in node.paths]
+        columns.append(node_data)
+
+    output = {
+        "query_column": result.query_column,
+        "direction": result.direction,
+        "count": len(result),
+        "columns": columns,
+    }
+    print(json.dumps(output, indent=2))
+
+
+def _format_query_result_csv(result) -> None:
+    """Format query result as CSV."""
+    print(
+        "identifier,table,column,hops,output_column,is_root,is_leaf,paths,file_path,query_index"
+    )
+    for node in result.related_columns:
+        file_path = node.file_path.replace('"', '""') if node.file_path else ""
+        # Format paths as semicolon-separated arrow strings
+        paths_str = (
+            ";".join(p.to_arrow_string() for p in node.paths) if node.paths else ""
+        )
+        paths_str = paths_str.replace('"', '""')
+
+        print(
+            f'"{node.identifier}","{node.table or ""}","{node.column or ""}",'
+            f'{node.hops},"{node.output_column}",'
+            f"{'true' if node.is_root else 'false'},"
+            f"{'true' if node.is_leaf else 'false'},"
+            f'"{paths_str}","{file_path}",{node.query_index}'
+        )
+
+
+if __name__ == "__main__":
+    app()