Remove init

Author: penguine-ip

a.py

@@ -0,0 +1,29 @@
+import asyncio
+from deepeval import evaluate
+from deepeval.metrics import AnswerRelevancyMetric
+from deepeval.test_case import LLMTestCase
+
+# Define the test cases
+test_case1 = LLMTestCase(
+    input="The chicken crossed the road, why?",
+    actual_output="Because he felt like it",
+)
+test_case2 = LLMTestCase(
+    input="Knock knock, who's there?", actual_output="The chicken"
+)
+
+# Define the metric
+metric1 = AnswerRelevancyMetric(verbose_mode=False)
+
+# # Asynchronous function to measure relevancy concurrently
+# async def measure_relevancy():
+#     await metric1.a_measure(test_case1, _show_indicator=False)
+#     print(metric1.statements)
+#     print("All measurements are done.")
+
+# # Run the asynchronous function and print after completion
+# asyncio.run(measure_relevancy())
+# print("This is printed after all asynchronous operations are complete.")
+
+
+evaluate([test_case1], [metric1])

deepeval/evaluate.py

@@ -308,6 +308,7 @@ async def a_execute_test_cases(
 
             for metric in metrics:
                 metric_metadata = create_metric_metadata(metric)
+                print(metric_metadata)
                 if isinstance(test_case, ConversationalTestCase):
                     # index hardcoded as the last message for now
                     api_test_case.update(

deepeval/metrics/answer_relevancy/answer_relevancy.py

@@ -56,7 +56,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._statements: ContextVar[Optional[List[str]]] = ContextVar(
             generate_uuid(), default=None
         )
@@ -129,7 +128,7 @@ async def a_measure(
             self.success = self.score >= self.threshold
             if self.verbose_mode:
                 print(
-                    f"statements: {self.statements}\nverdicts: {self.verdicts}\nscore: {self.score}, success: {self.success}\n"
+                    f"statements: {self.statements}\nverdicts: {self.verdicts}\nscore: {self.score}, success: {self.success}, reason: {self.reason}\n"
                 )
             return self.score
 

deepeval/metrics/base_metric.py

@@ -16,22 +16,21 @@ class BaseMetric:
     include_reason: bool = False
     evaluation_cost: Optional[float] = None
 
-    def __init__(self):
-        self._score: ContextVar[Optional[float]] = ContextVar(
-            generate_uuid(), default=None
-        )
-        self._score_breakdown: ContextVar[Optional[Dict]] = ContextVar(
-            generate_uuid(), default=None
-        )
-        self._reason: ContextVar[Optional[str]] = ContextVar(
-            generate_uuid(), default=None
-        )
-        self._success: ContextVar[Optional[bool]] = ContextVar(
-            generate_uuid(), default=None
-        )
-        self._error: ContextVar[Optional[str]] = ContextVar(
-            generate_uuid(), default=None
-        )
+    _score: ContextVar[Optional[float]] = ContextVar(
+        generate_uuid(), default=None
+    )
+    _score_breakdown: ContextVar[Optional[Dict]] = ContextVar(
+        generate_uuid(), default=None
+    )
+    _reason: ContextVar[Optional[str]] = ContextVar(
+        generate_uuid(), default=None
+    )
+    _success: ContextVar[Optional[bool]] = ContextVar(
+        generate_uuid(), default=None
+    )
+    _error: ContextVar[Optional[str]] = ContextVar(
+        generate_uuid(), default=None
+    )
 
     @property
     def score(self) -> Optional[float]:

deepeval/metrics/bias/bias.py

@@ -58,7 +58,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._opinions: ContextVar[Optional[List[str]]] = ContextVar(
             generate_uuid(), default=None
         )

deepeval/metrics/contextual_precision/contextual_precision.py

@@ -53,7 +53,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._verdicts: ContextVar[
             Optional[List[ContextualPrecisionVerdict]]
         ] = ContextVar(generate_uuid(), default=None)

deepeval/metrics/contextual_recall/contextual_recall.py

@@ -50,7 +50,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._verdicts: ContextVar[Optional[List[ContextualRecallVerdict]]] = (
             ContextVar(generate_uuid(), default=None)
         )

deepeval/metrics/contextual_relevancy/contextual_relevancy.py

@@ -52,7 +52,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._verdicts: ContextVar[
             Optional[List[ContextualRelevancyVerdict]]
         ] = ContextVar(generate_uuid(), default=None)

deepeval/metrics/faithfulness/faithfulness.py

@@ -67,7 +67,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._truths: ContextVar[Optional[List[str]]] = ContextVar(
             generate_uuid(), default=None
         )

deepeval/metrics/g_eval/g_eval.py

@@ -75,7 +75,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._evaluation_steps: ContextVar[Optional[List[str]]] = ContextVar(
             generate_uuid(), default=None
         )

deepeval/metrics/hallucination/hallucination.py

@@ -50,7 +50,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._verdicts: ContextVar[Optional[List[HallucinationVerdict]]] = (
             ContextVar(generate_uuid(), default=None)
         )

deepeval/metrics/knowledge_retention/knowledge_retention.py

@@ -52,7 +52,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._knowledges: ContextVar[Optional[List[Knowledge]]] = ContextVar(
             generate_uuid(), default=None
         )

deepeval/metrics/summarization/summarization.py

@@ -112,7 +112,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._truths: ContextVar[Optional[List[str]]] = ContextVar(
             generate_uuid(), default=None
         )

deepeval/metrics/toxicity/toxicity.py

@@ -58,7 +58,6 @@ def __init__(
         strict_mode: bool = False,
         verbose_mode: bool = False,
     ):
-        super().__init__()
         self._opinions: ContextVar[Optional[List[str]]] = ContextVar(
             generate_uuid(), default=None
         )

deepeval/test_run/test_run.py

@@ -511,6 +511,7 @@ def post_test_run(self, test_run: TestRun):
             except AttributeError:
                 # Pydantic version below 2.0
                 body = test_run.dict(by_alias=True, exclude_none=True)
+
             api = Api()
             result = api.post_request(
                 endpoint=Endpoints.TEST_RUN_ENDPOINT.value,

tests/test_everything.py

@@ -82,7 +82,7 @@
 strict_mode = False
 
 
-# @pytest.mark.skip(reason="openai is expensive")
+@pytest.mark.skip(reason="openai is expensive")
 def test_everything():
     metric1 = AnswerRelevancyMetric(
         threshold=0.1,