zedda.validate()

zedda.validate(
    data,
    rules: dict[str, dict[str, Any]],
    profile=None,
    fail_on_error: bool = False,
) -> ValidationReport

Validate a dataset against a declarative data quality contract. Evaluates column existence, null thresholds, value ranges, allowed types, and cardinality constraints.

# Arguments

Argument	Type	Default	Description
data	str / Path / DataFrame	(required)	Input dataset path or in-memory DataFrame
rules	dict[str, dict[str, Any]]	(required)	Declarative column validation rules dictionary
profile	DatasetProfile or None	None	Pre-computed profile. If None, zd.scan(data) is run automatically
fail_on_error	bool	False	If True, raises ZeddaError if any critical rule fails

# Supported Rule Specifications

Rules are defined as a mapping from column name to constraint dictionary:

rules = {
    "age": {
        "max_null_pct": 5.0,        # Maximum allowed null percentage
        "min_val": 0,                # Minimum allowed numeric value
        "max_val": 120,              # Maximum allowed numeric value
        "allowed_types": ["int", "float"], # Allowed data types
    },
    "email": {
        "max_null_pct": 0.0,
        "min_unique": 10,            # Minimum unique values
    },
}

Rule	Type	Description
max_null_pct	float	Maximum null percentage allowed (0.0 to 100.0)
min_val	float / int	Lower bound for numeric columns
max_val	float / int	Upper bound for numeric columns
allowed_types	list[str]	Allowed type strings (e.g. ["int", "float", "str", "bool"])
min_unique	int	Minimum unique count (via HyperLogLog / exact count)
max_unique	int	Maximum unique count

# Returns

A ValidationReport object containing:

• is_valid (bool): True if all rules passed with zero breaches.
• passed_rules (int): Count of rules satisfied.
• failed_rules (int): Count of rules breached.
• columns (list): Per-column breach breakdown.
• summary() (str): Multi-line formatted terminal summary.

# Raises

ZeddaError if fail_on_error=True and one or more validation rules fail.

# Example

import zedda as zd

contract = {
    "fare": {"min_val": 0.0, "max_null_pct": 1.0},
    "survived": {"allowed_types": ["int", "bool"]},
}

report = zd.validate("titanic.csv", rules=contract)

if not report.is_valid:
    print(f"Validation failed: {report.failed_rules} breaches detected")
    print(report.summary())

# CLI Equivalent

zedda validate data.csv --rules contract.json --fail