Merge pull request confident-ai#138 from confident-ai/feature/refacto…

…r-ragas

update ragas

deepeval/_version.py

@@ -1 +1 @@
-__version__: str = "0.17.4"
+__version__: str = "0.17.5"

deepeval/metrics/ragas_metric.py

@@ -7,7 +7,7 @@
 from typing import List
 
 
-class ContextualRelevancyRagasMetric(Metric):
+class RagasContextualRelevancyMetric(Metric):
     """This metric checks the contextual relevancy using Ragas"""
 
     def __init__(
@@ -64,10 +64,10 @@ def is_successful(self):
 
     @property
     def __name__(self):
-        return "Contextual Relevancy Ragas Score"
+        return "Ragas Contextual Relevancy Score"
 
 
-class AnswerRelevancyRagasMetric(Metric):
+class RagasAnswerRelevancyMetric(Metric):
     """This metric checks the answer relevancy using Ragas"""
 
     def __init__(
@@ -117,10 +117,10 @@ def is_successful(self):
 
     @property
     def __name__(self):
-        return "Answer Relevancy Ragas Score"
+        return "Ragas Answer Relevancy Score"
 
 
-class FaithfulnessRagasMetric(Metric):
+class RagasFaithfulnessMetric(Metric):
     def __init__(
         self,
         minimum_score: float = 0.3,
@@ -168,10 +168,10 @@ def is_successful(self):
 
     @property
     def __name__(self):
-        return "Faithfulness Ragas Score"
+        return "Ragas Faithfulness Score"
 
 
-class ContextRecallRagasMetric(Metric):
+class RagasContextRecallMetric(Metric):
     """This metric checks the context recall using Ragas"""
 
     def __init__(
@@ -221,10 +221,10 @@ def is_successful(self):
 
     @property
     def __name__(self):
-        return "Context Recall Ragas Score"
+        return "Ragas Context Recall Score"
 
 
-class HarmfulnessRagasMetric(Metric):
+class RagasHarmfulnessMetric(Metric):
     """This metric checks the harmfulness using Ragas"""
 
     def __init__(
@@ -274,7 +274,7 @@ def is_successful(self):
 
     @property
     def __name__(self):
-        return "Harmfulness Ragas Score"
+        return "Ragas Harmfulness Score"
 
 
 class RagasMetric(Metric):
@@ -288,11 +288,11 @@ def __init__(
         self.minimum_score = minimum_score
         if metrics is None:
             self.metrics = [
-                HarmfulnessRagasMetric,
-                ContextRecallRagasMetric,
-                FaithfulnessRagasMetric,
-                AnswerRelevancyRagasMetric,
-                ContextualRelevancyRagasMetric,
+                RagasHarmfulnessMetric,
+                RagasContextRecallMetric,
+                RagasFaithfulnessMetric,
+                RagasAnswerRelevancyMetric,
+                RagasContextualRelevancyMetric,
             ]
         else:
             self.metrics = metrics

docs/docs/measuring_llm_performance/ragas_score.md

@@ -43,3 +43,37 @@ def test_overall_score():
     )
 ```
 
+
+## Individual Metrics
+
+You can call any of the following methodologies to use specific Ragas metrics. Below is an example of how you can do just that!
+
+```python
+from deepeval.metrics.ragas_metric import RagasContextualRelevancyMetric
+from deepeval.metrics.ragas_metric import RagasAnswerRelevancyMetric
+from deepeval.metrics.ragas_metric import RagasFaithfulnessMetric
+from deepeval.metrics.ragas_metric import RagasContextRecallMetric
+from deepeval.metrics.ragas_metric import RagasHarmfulnessMetric
+
+def test_individual_metrics():
+    test_case = LLMTestCase(
+        query=query,
+        output=output,
+        expected_output=expected_output,
+        context=context,
+    )
+    metrics = [
+        RagasContextualRelevancyMetric(),
+        RagasAnswerRelevancyMetric(),
+        RagasFaithfulnessMetric(),
+        RagasContextRecallMetric(),
+        RagasHarmfulnessMetric(),
+    ]
+    for metric in metrics:
+        score = metric.measure(test_case)
+        print(f"{metric.__name__}: {score}")
+```
+
+This will print the individual scores for each metric.
+
+