Make evaluation name optional (#66)

* Make evaluation name optional

* Adding tests to check empty evaluation name

* Adjust import

* Address comments

---------

Co-authored-by: Ramon Marques <ramon.marques@toptal.com>

Author: ramoncelestino

aimon/decorators/evaluate.py

@@ -1,5 +1,5 @@
 from functools import wraps
-
+from datetime import datetime
 from aimon import Client
 import inspect
 import warnings
@@ -144,8 +144,8 @@ def evaluate(
         application_name,
         model_name,
         dataset_collection_name, 
-        evaluation_name, 
-        headers, 
+        evaluation_name=None,
+        headers=None,
         api_key=None,
         aimon_client=None,
         config=None
@@ -222,6 +222,11 @@ def evaluate(
     application = Application(name=application_name, stage="evaluation")
     model = Model(name=model_name, model_type="text")
 
+    # Auto-generate evaluation name if not provided
+    if not evaluation_name:
+        timestamp = datetime.utcnow().strftime("%Y%m%dT%H%M%S")
+        evaluation_name = f"{application_name}-{model_name}-{timestamp}"
+
     # Validata headers to be non-empty and contain atleast the context_docs column
     if not headers:
         raise ValueError("Headers must be a non-empty list")

tests/test_evaluate.py

@@ -531,4 +531,46 @@ def test_evaluate_with_custom_client(self):
 
         except Exception as e:
             self.log_info("Test error", str(e))
-            raise 
+            raise
+
+    def test_evaluate_without_evaluation_name(self):
+        """Test the evaluate function when no evaluation name is provided."""
+        if not self.api_key:
+            pytest.skip("AIMON_API_KEY environment variable not set")
+
+        try:
+            # Create test data
+            test_data = self.create_test_data()
+
+            headers = ["context_docs", "user_query", "output", "prompt", "task_definition"]
+            config = {'hallucination': {'detector_name': 'default'}}
+
+            self.log_info("Starting evaluate test without evaluation_name", {
+                "Application": self.app_name,
+                "Model": self.model_name,
+                "Collection": self.collection_name,
+                "Headers": headers,
+                "Config": config
+            })
+
+            # Call evaluate without providing evaluation_name
+            results = evaluate(
+                application_name=self.app_name,
+                model_name=self.model_name,
+                dataset_collection_name=self.collection_name,
+                headers=headers,  # evaluation_name omitted
+                api_key=self.api_key,
+                config=config
+            )
+
+            assert len(results) == 2
+            assert isinstance(results[0], EvaluateResponse)
+            assert results[0].output in ["Paris is the capital of France.", "Python is a versatile programming language."]
+            assert hasattr(results[0].response, 'status')
+            assert results[0].response.status == 200
+
+            self.log_info("Test completed successfully", "Auto-generated evaluation_name handled correctly")
+
+        except Exception as e:
+            self.log_info("Test error", str(e))
+            raise