# fugue-project/fugue

A unified interface for distributed computing. Fugue executes SQL, Python, Pandas, and Polars code on Spark, Dask and Ray without any rewrites.

Repository: https://github.com/fugue-project/fugue
Canonical: https://ross.abutalabs.com/products/fugue
Homepage: https://fugue-tutorials.readthedocs.io/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: spark, dask, data-practitioners, machine-learning, distributed-systems, distributed-computing, distributed, sql, pandas
Last push: 2026-05-19T21:49:03+00:00

## Health v2 (maintenance only)
Score: 82/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 83, release rhythm 71, longevity 100
- inputs: {"age_days": 2353, "days_push": 106, "days_rel": 194, "gap_med": 28, "n_releases_24m": 6}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2169, forks 103 (observed 2026-08-28T04:06:21.561357+00:00)

## What it is
Fugue is a Python library providing a unified interface for distributed computing, letting users run Python, Pandas, Polars, and SQL code on Spark, Dask, and Ray backends with minimal rewrites. It includes FugueSQL, an enhanced SQL interface that can invoke Python code for end-to-end data workflows.

## Use cases
- scale pandas code to spark dask or ray
- parallelize python functions across a cluster
- run sql workflows on top of pandas and spark dataframes
- port local data pipelines to distributed backends without rewrites
- bridge local testing and large-scale distributed execution
- invoke python functions from sql queries

## When to choose
- you have pandas or python code that needs to scale to Spark, Dask, or Ray
- you want a SQL interface that works across multiple dataframe engines
- you want to avoid learning framework-specific APIs for each distributed backend
- you need to test logic locally and run the same code at scale

## When to avoid
- your data fits comfortably in memory on a single machine
- you want a drop-in pandas replacement like Modin or PySpark Pandas
- you are fully committed to one engine and its native APIs
- you need dbt-style SQL transformation management for warehouses

## Facets
- artifact type: library
- maturity: active
- function: etl, data-science, workflow-automation
- domain: big-data, microservices, data-science
- platform: python, cross-platform
- tags: distributed-computing, spark, dask, ray, pandas, polars, fuguesql, dataframes, sql, data-engineering

## Member repositories
- fugue-project/fugue (main) score 82

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:21.561357+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:49:26.767214+00:00, confidence not recorded.
  - readme: https://github.com/fugue-project/fugue (fetched 2026-08-28T04:06:21.561357+00:00, sha 31e049efc1ea)
  - homepage: https://fugue-tutorials.readthedocs.io/ (fetched 2026-08-29T10:29:29.202610+00:00, sha c8e45ef96907)
  - registry_pypi: https://pypi.org/pypi/fugue/json (fetched 2026-08-29T10:29:29.212228+00:00, sha 5a1231364ef5)
- Data as of 2026-08-30T08:39:29.467469+00:00.