arnio

Data trust for Python.

Validate, clean, and profile DataFrames before everything else.

What is Arnio?

Arnio is the data trust layer for Python. It answers one question:

"Is this data trustworthy?"

It answers through three verbs: validate, clean, and profile.

Think of it as Pydantic for DataFrames.

Install

pip install arnio

Quick Start

import arnio as ar
import pandas as pd

df = pd.DataFrame({
    "email": ["alice@example.com", "not-an-email", None],
    "age": [25, -5, 200],
    "name": ["Alice", "Bob", "Charlie"],
})

# Profile — instant data quality overview
report = ar.profile(df)
print(report.quality_score)  # 0–100

# Validate — check against a schema
schema = ar.Schema({
    "email": ar.Email(nullable=False),
    "age": ar.Int(min=0, max=150),
    "name": ar.String(min_length=1),
})
result = ar.validate(df, schema)
print(result.passed)   # False
print(result.issues)   # Structured list of issues

# Clean — declarative cleaning pipeline
cleaned = ar.clean(df, [
    "strip_whitespace",
    "drop_duplicates",
    ("normalize_case", {"case": "lower"}),
])

# Suggest — intelligent cleaning suggestions
suggestions = ar.suggest(df)

Three Core Verbs

Verb	Function	Purpose
Validate	`ar.validate(df, schema)`	Check if data matches a contract
Clean	`ar.clean(df, steps)`	Apply declarative cleaning operations
Profile	`ar.profile(df)`	Measure data quality with scores and metrics

Schema Definition

Two ways to define the same schema:

# Dict-based (dynamic, config-driven)
schema = ar.Schema({
    "email": ar.Email(nullable=False),
    "age": ar.Int(min=0, max=150),
    "name": ar.String(min_length=1),
})

# Class-based (IDE-friendly, inheritable)
class Customers(ar.Schema):
    email = ar.Email(nullable=False)
    age = ar.Int(min=0, max=150)
    name = ar.String(min_length=1)

Available Field Types

Type	Description
`ar.Int`	Integer with optional min/max
`ar.Float`	Float with optional min/max
`ar.String`	String with optional length/pattern
`ar.Bool`	Boolean
`ar.Date`	Date string with format validation
`ar.DateTime`	DateTime string with format validation
`ar.Email`	Email address
`ar.URL`	HTTP/HTTPS URL
`ar.PhoneNumber`	Phone number
`ar.IPAddress`	IPv4 or IPv6 address
`ar.UUID`	UUID string
`ar.Regex`	Custom regex pattern

Cleaning Pipeline

# One-shot cleaning
cleaned = ar.clean(df, [
    "strip_whitespace",
    "drop_duplicates",
    ("fill_nulls", {"column": "category", "value": "unknown"}),
    "slugify_column_names",
])

# Reusable pipeline
pipe = ar.Pipeline([
    "strip_whitespace",
    "drop_duplicates",
    ("normalize_case", {"case": "lower"}),
])
cleaned = pipe.run(df)

# Save/load for version control
yaml_str = pipe.to_yaml()
pipe = ar.Pipeline.from_yaml(yaml_str)

Quality Gates (CI/CD)

# In a test file or CI script
ar.check(df, schema)  # Raises ar.ValidationError on failure

pandas Accessor

import arnio  # Registers the accessor

df.arnio.profile()
df.arnio.validate(schema)
df.arnio.clean(["strip_whitespace"])
df.arnio.suggest()
df.arnio.is_valid(schema)  # Returns bool

Works With

pandas DataFrames (primary)
dict / list of dicts (no pandas import needed for simple cases)
Polars DataFrames (v2.1)

What Arnio Does NOT Do

Arnio stays in its lane. It does not:

Query or filter data (use pandas/Polars)
Perform analytics (use pandas/Polars/DuckDB)
Do feature engineering (use scikit-learn)
Create visualizations (use matplotlib/Plotly)
Manage file formats (use PyArrow)

License

MIT — Anish Raj

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

arnio-2.0.0.tar.gz (39.1 MB view details)

Uploaded Jul 15, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

arnio-2.0.0-py3-none-any.whl (45.9 kB view details)

Uploaded Jul 15, 2026 Python 3

File details

Details for the file arnio-2.0.0.tar.gz.

File metadata

Download URL: arnio-2.0.0.tar.gz
Upload date: Jul 15, 2026
Size: 39.1 MB
Tags: Source
Uploaded using Trusted Publishing? No
Uploaded via: uv/0.9.28 {"installer":{"name":"uv","version":"0.9.28","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":null,"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

File hashes

Hashes for arnio-2.0.0.tar.gz
Algorithm	Hash digest
SHA256	`cc509758b7b90f431ffe14622b13bccb0f2a1a83e50f7ab1c318f08014800938`
MD5	`1b4a2823ce36f2d1ed909d9614ce78fc`
BLAKE2b-256	`2b44b8d17ffa7db332813f953b22c1c709e7e8339e40e4b78fb41b4f86d5dcf4`

See more details on using hashes here.

File details

Details for the file arnio-2.0.0-py3-none-any.whl.

File metadata

Download URL: arnio-2.0.0-py3-none-any.whl
Upload date: Jul 15, 2026
Size: 45.9 kB
Tags: Python 3
Uploaded using Trusted Publishing? No
Uploaded via: uv/0.9.28 {"installer":{"name":"uv","version":"0.9.28","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":null,"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

File hashes

Hashes for arnio-2.0.0-py3-none-any.whl
Algorithm	Hash digest
SHA256	`733a00015b7c65cc0bf26b546a3cc463bb2b8097795c763edb16db07da7ed766`
MD5	`a377fff355e527f5b9dd0dd0a622fc95`
BLAKE2b-256	`cde0fab08dabda88b8619c3e1e6c9bfe3f486ef0725687465c35fd4b7e95319e`