Skip to content
ixtract

Quick Start

Install

Section titled “Install”
Terminal window
pip install ixtract

Python 3.12+ required. Tested on Ubuntu and macOS.

Connect to PostgreSQL

Section titled “Connect to PostgreSQL”
from ixtract import plan, execute, ExtractionIntent


intent = ExtractionIntent(
    source_type="postgresql",
    source_config={
        "host":     "localhost",
        "port":     5432,
        "database": "mydb",
        "user":     "app",
        "password": "secret",
    },
    object_name="orders",
)

Plan and execute

Section titled “Plan and execute”
result = plan(intent)


if result.is_safe:
    execution = execute(result)
    print(f"{execution.rows_extracted:,} rows in {execution.duration_seconds:.1f}s")

plan() profiles the source, consults run history, and returns an ExecutionPlan. Nothing is extracted until you call execute().

CLI equivalent

Section titled “CLI equivalent”
Terminal window
ixtract execute orders \
  --database mydb --user app --password secret

What happens on first run

Section titled “What happens on first run”

On the first run, ixtract has no history. It profiles the table, estimates row count and PK distribution, and selects a conservative worker count (default: 4). After the run completes, it stores the result in a local SQLite state store.

On subsequent runs, the controller uses that history to converge toward an optimal worker count.

Check the plan before running

Section titled “Check the plan before running”
Terminal window
ixtract plan orders --database mydb --user app --password secret --standard

--standard shows the full planning output including worker resolution, chunk boundaries, and verdict.

Inspect history

Section titled “Inspect history”
Terminal window
ixtract history orders

Shows all past runs: workers, throughput, duration, and diagnosis.

Next steps

Section titled “Next steps”