Fill Rate Accuracy

Fill Rate Accuracy measures how well a "got" object matches an "expected" object in terms of field completeness states (filled vs missing). Unlike Similarity, it focuses on whether fields are present, not their actual values.

Concept

When comparing two objects: - Fill Rate Accuracy = 1.0 if both have the same state for a field (both filled or both missing) - Fill Rate Accuracy = 0.0 if states differ (one filled, one missing)

This is useful for validating that your data generation pipeline produces the right "shape" of data, regardless of the actual values.

Basic Usage

from cobjectric import BaseModel

class Person(BaseModel):
    name: str
    age: int
    email: str

# Generated data (missing email)
got = Person.from_dict({"name": "John", "age": 30})

# Expected data (has email)
expected = Person.from_dict({
    "name": "Jane",
    "age": 25,
    "email": "[email protected]"
})

accuracy = got.compute_fill_rate_accuracy(expected)
print(accuracy.fields.name.value)   # 1.0 (both filled)
print(accuracy.fields.age.value)    # 1.0 (both filled)
print(accuracy.fields.email.value)  # 0.0 (got missing, expected filled)
print(accuracy.mean())              # 0.667 (2 out of 3 match)

Note: The actual values of name and age don't matter. Fill Rate Accuracy only cares about field presence/absence.

Key Differences

Fill Rate Accuracy vs Similarity

Aspect	Fill Rate Accuracy	Similarity
Focuses on	Field state (present/missing)	Field values
Use case	Validating data shape	Validating data quality
Example	Checks if email field exists	Checks if email value is correct

Example Comparison

got = Person.from_dict({"name": "John Doe", "age": 30, "email": "[email protected]"})
expected = Person.from_dict({"name": "jane doe", "age": 30, "email": "[email protected]"})

# Fill Rate Accuracy (comparing states)
accuracy = got.compute_fill_rate_accuracy(expected)
print(accuracy.fields.name.value)     # 1.0 (both fields filled)

# Similarity (comparing values)
similarity = got.compute_similarity(expected)
print(similarity.fields.name.value)   # ~0.92 (fuzzy match: "john" vs "jane")

Use Cases

1. Data Pipeline Validation: Ensure your data generation produces complete objects
2. Quality Control: Verify that all required fields are populated
3. Schema Conformance: Check that generated data matches the expected structure
4. Testing: Assert that generated data has the correct shape

Advanced Usage

You can combine Fill Rate Accuracy with other metrics:

# Check data shape
accuracy = got.compute_fill_rate_accuracy(expected)
if accuracy.mean() < 0.8:
    print("Warning: Data shape doesn't match!")

# Then check data quality
similarity = got.compute_similarity(expected)
if similarity.mean() < 0.9:
    print("Warning: Data values don't match!")

See Similarity for value-based comparison and Field Specifications for custom validation logic.

API Reference

See the API Reference for the complete Fill Rate Accuracy result classes and methods.