# approximatelabs/sketch

AI code-writing assistant that understands data content

Repository: https://github.com/approximatelabs/sketch
Canonical: https://ross.abutalabs.com/products/approximatelabs-sketch
Language: Python
License: MIT
License Family: permissive
Topics: ai, codex, copilot, data, data-science, dataframe, datasketches, df, ds, gpt3, pandas, sketches, tabular-data, datasketch, lambdaprompt, python
Last push: 2024-02-08T15:15:58+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1512, "days_push": 937, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2284, forks 121 (observed 2026-08-28T04:06:33.983730+00:00)

## What it is
Sketch is an AI code-writing assistant for pandas users that understands the context of your dataframe content to generate relevant suggestions. It adds a .sketch extension to any pandas dataframe providing question-answering, code generation, and data generation prompts.

## Use cases
- ask questions about my pandas dataframe contents
- generate pandas code from natural language
- clean and mask data columns with AI
- generate new features from text columns in a dataframe
- identify PII columns in my dataset
- get plot code for my dataframe

## When to choose
- you work with pandas dataframes and want AI suggestions aware of your actual data
- you want an AI assistant without installing an IDE plugin
- you need quick data cataloging, cleaning, or feature-generation help in notebooks

## When to avoid
- you need a fully local/offline assistant since it relies on OpenAI and a hosted prompt service
- you don't use pandas or tabular data
- you need actively developed tooling with recent updates

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, nlp, data-science, prompt-engineering
- domain: data-science, large-language-models, developer-tools
- platform: python, cross-platform
- tags: pandas, dataframes, ai-code-assistant, gpt3, tabular-data, data-cataloging, data-cleaning, openai

## Member repositories
- approximatelabs/sketch (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:33.983730+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 2026-08-30T02:41:05.659133+00:00, confidence not recorded.
  - readme: https://github.com/approximatelabs/sketch (fetched 2026-08-28T04:06:33.983730+00:00, sha 6c2b06b06458)
  - registry_pypi: https://pypi.org/pypi/sketch/json (fetched 2026-08-29T10:21:19.119521+00:00, sha e085e6f5c2a3)
- Data as of 2026-08-30T08:39:29.467469+00:00.