API Overview

API Overview

This page maps the main public API surface. For exact signatures, use the repository's API_REFERENCE.md.

Core Frame

ArFrame is Arnio's in-memory frame wrapper around the C++ backend.

Common properties and methods:

• columns
• dtypes
• shape
• is_empty
• memory_usage()
• preview()
• select_columns()
• select_dtypes()
• to_dict()

Input and Output

Function	Purpose
read_csv	Load CSV, TSV, or TXT data into an ArFrame.
read_csv_chunked	Iterate over large CSV input in chunks.
scan_csv	Infer schema without loading the whole dataset.
write_csv	Write an ArFrame to CSV, TSV, or TXT.
read_jsonl	Read JSON Lines data into an ArFrame.
write_parquet	Write parquet output when optional dependencies are available.
sniff_delimiter	Detect a CSV delimiter from a sample.

Conversion

Function	Purpose
from_pandas	Convert a pandas DataFrame into an ArFrame.
to_pandas	Convert an ArFrame into a pandas DataFrame.
to_arrow	Convert an ArFrame into a pyarrow Table.

Cleaning

Common cleaning helpers:

• strip_whitespace
• normalize_case
• normalize_unicode
• fill_nulls
• drop_nulls
• keep_rows_with_nulls
• drop_duplicates
• drop_columns
• drop_constant_columns
• drop_empty_columns
• trim_column_names
• rename_columns
• cast_types
• filter_rows
• replace_values
• round_numeric_columns
• clip_numeric
• winsorize_outliers
• combine_columns
• safe_divide_columns
• coalesce_columns
• standardize_missing_tokens

Pipeline

Function	Purpose
pipeline	Apply ordered cleaning steps.
register_step	Register a custom Python pipeline step.
list_steps	Inspect available pipeline steps.
reset_steps	Reset custom step state.

Data Quality

API	Purpose
profile	Build a DataQualityReport.
suggest_cleaning	Suggest cleaning steps from a frame or report.
auto_clean	Profile and apply safe cleaning actions.
compare_profiles	Compare two quality profiles.
check_quality_gates	Gate quality drift in CI or monitoring.

Schema Validation

Core classes:

• Schema
• Field
• ValidationResult
• ValidationIssue
• SchemaDiff

Field factories:

• Int64
• Float64
• String
• Bool
• Email
• URL
• PhoneNumber
• CountryCode
• CurrencyCode
• Date
• DateTime
• Regex
• Custom

Helpers:

• validate
• diff_schema
• register_validator
• schema_to_dict
• schema_to_yaml

Exceptions

• ArnioError
• CsvReadError
• JsonlReadError
• TypeCastError
• UnknownStepError
• PipelineStepError