phlo 0.7.9

Lakehouse platform

Modern data lakehouse platform. Plugin-driven. Storage-agnostic.

Features

• Decorator-driven development — @phlo.ingestion and @phlo.quality replace hundreds of lines of boilerplate
• Write-Audit-Publish pattern — Git-like branching with automatic quality gates and promotion
• Type-safe data quality — Pandera schemas enforce validation at ingestion time
• Plugin architecture — 12 plugin types: sources, quality, ingestion, transforms, services, hooks, catalogs, assets, resources, orchestrators, and CLI commands
• Storage-agnostic — Iceberg, Delta, or bring-your-own via table-format plugins
• Observatory UI — Web-based data exploration, lineage, and monitoring
• Observability — OpenTelemetry traces, metrics, and logs via phlo-otel; Grafana/Prometheus/Loki stack
• Production-ready — Auto-publishing, configurable merge strategies, freshness policies, data migrations

What It Looks Like

import phlo

@phlo.ingestion(
    table_name="events",
    unique_key="id",
    validation_schema=EventSchema,
    group="api",
    cron="0 */1 * * *",
    freshness_hours=(1, 24),
)
def api_events(partition_date: str):
    return rest_api(...)  # Any DLT source


@phlo.quality(
    table="bronze.events",
    checks=[
        NullCheck(columns=["id", "timestamp"]),
        RangeCheck(column="value", min_value=0, max_value=100),
        UniqueCheck(columns=["id"]),
        FreshnessCheck(column="timestamp", max_age_hours=24),
    ],
)
def events_quality():
    pass

Prerequisites

• uv — Python package manager
• Docker — Container runtime

Quick Start

# Install with default plugins
uv pip install phlo[defaults]

# Initialize a new project
phlo init my-project
cd my-project

# Start services and materialize
phlo services start
phlo materialize --select "dlt_glucose_entries+"

Documentation

Full documentation source lives under docs/index.md. The Fumadocs site now builds from docs-app/ against that same Markdown tree:

npm --prefix docs-app install
npm --prefix docs-app run dev

The docs app serves locally on http://localhost:3101.

Primary entry points:

• Installation Guide
• Quickstart Guide
• Core Concepts
• Developer Guide
• Plugin Development
• Workflow Development
• CLI Reference
• Configuration Reference
• Operations Guide

Development

uv pip install -e .    # Install Phlo in dev mode
make check             # Lint, format, typecheck, and test (parallel)

# Services
phlo services start    # Start infrastructure
phlo services stop     # Stop services
phlo services logs -f  # View logs

# Individual gates
uv run ruff check .    # Lint
uv run ruff format .   # Format
uv run ty check        # Typecheck
uv run pytest          # Test

Architecture

Phlo is a monorepo of composable packages — install only what you need:

Layer	Packages
Orchestration	phlo-dagster
Ingestion	phlo-dlt
Quality	phlo-pandera
Transforms	phlo-dbt
Table formats	phlo-iceberg, phlo-delta, phlo-clickhouse
Infrastructure	phlo-traefik, phlo-postgres
Storage	phlo-minio
Catalog	phlo-nessie, phlo-openmetadata
Query	phlo-trino
Observability	phlo-otel, phlo-clickstack, phlo-grafana, phlo-prometheus, phlo-loki, phlo-alloy
UI	phlo-observatory, phlo-pgweb, phlo-superset
API	phlo-api, phlo-hasura, phlo-postgrest
Dev/Test	phlo-testing