MetricResult

MetricResult object to store the result of a metric

---

source

MetricResult

 MetricResult (result:Any, reason:Optional[str]=None,
               traces:Optional[Dict[str,Any]]=None)

*Class to hold the result of a metric evaluation.

This class behaves like its underlying result value but still provides access to additional metadata like reasoning.

Works with: - DiscreteMetrics (string results) - NumericMetrics (float/int results) - RankingMetrics (list results)*

Example Usage

metric_result = MetricResult(result=42, reason="This is a test")
print(metric_result)
print(metric_result.reason)

### Example with Numeric Operations
num_result1 = MetricResult(result=5.0)
num_result2 = MetricResult(result=3.0)
print(num_result1 + num_result2)  # 8.0


### Example with String Operations
str_result = MetricResult(result="low")
print(str_result.upper())  # "LOW"

## Example with List Operations
list_result = MetricResult(result=[1, 2, 3])
print(list_result[1:])  # 2

42
This is a test
8.0
LOW
[2, 3]

now lets make it Pydantic compatible also

---

source

MetricResult.__get_pydantic_core_schema__

 MetricResult.__get_pydantic_core_schema__ (_source_type:Any,
                                            _handler:pydantic.annotated_ha
                                            ndlers.GetCoreSchemaHandler)

*Generate a Pydantic core schema for MetricResult.

This custom schema handles different serialization behaviors: - For model_dump(): Returns the original MetricResult instance - For model_dump_json(): Converts to a JSON-compatible dict using json*

---

source

MetricResult.__json__

 MetricResult.__json__ ()

*Return data for JSON serialization.

This method is used by json.dumps and other JSON serializers to convert MetricResult to a JSON-compatible format.*

---

source

MetricResult.validate

 MetricResult.validate (value:Any,
                        info:pydantic_core.core_schema.ValidationInfo)

Provide compatibility with older Pydantic versions.

from pydantic import BaseModel

class TestModel(BaseModel):
    response: str
    grade: MetricResult
    faithfulness: MetricResult

m = TestModel(response="test", grade=MetricResult(result=1, reason="test"), faithfulness=MetricResult(result=1, reason="test"))
m

TestModel(response='test', grade=1, faithfulness=1)

m.model_dump()

{'response': 'test', 'grade': 1, 'faithfulness': 1}

m.model_dump_json()

'{"response":"test","grade":{"result":1,"reason":"test"},"faithfulness":{"result":1,"reason":"test"}}'