31. Building a CLI Tool

Command-line tools turn scripts into shareable utilities. Python's ecosystem offers several frameworks; click hits the sweet spot of simplicity and capability.

Getting started with click:

#!/usr/bin/env python3
"""CLI tool for processing AI documents."""

import click
from pathlib import Path

@click.group()
@click.version_option(version="1.0.0")
def cli():
    """AI document processing toolkit."""
    pass

@cli.command()
@click.argument("input_file", type=click.Path(exists=True))
@click.option("--output", "-o", type=click.Path(), help="Output file path")
@click.option("--verbose", "-v", is_flag=True, help="Enable verbose output")
def process(input_file, output, verbose):
    """Process a document through the AI pipeline."""
    if verbose:
        click.echo(f"Processing: {input_file}")
    
    with open(input_file) as f:
        content = f.read()
    
    word_count = len(content.split())
    char_count = len(content)
    
    if verbose:
        click.echo(f"Words: {word_count}, Characters: {char_count}")
    
    result = f"Document: {Path(input_file).name}\nWords: {word_count}\n"
    
    if output:
        Path(output).write_text(result)
        click.echo(f"Wrote to {output}")
    else:
        click.echo(result)

@cli.command()
def status():
    """Check pipeline status."""
    click.echo("Pipeline: Ready")
    click.echo("Models: Loaded")

if __name__ == "__main__":
    cli()

Run it directly: python cli_tool.py process myfile.txt -v. Or install it for command-line access. Run pip install -e . from your project root after adding to pyproject.toml:

[project.scripts]
ai-pipeline = "pipeline.cli:cli"

Local verification checkpoint

Run the smallest example from this chapter in a local workspace and record the package version, runtime, data path, and observed output. If the result depends on model size, vector count, CPU/GPU backend, or available memory, note that constraint beside the exercise so the lesson remains reproducible.

Local verification checkpoint

Run the smallest example from this chapter in a local workspace and record the package version, runtime, data path, and observed output. If the result depends on model size, vector count, CPU/GPU backend, or available memory, note that constraint beside the exercise so the lesson remains reproducible.