judgeval 0.0.26__py3-none-any.whl → 0.0.28__py3-none-any.whl

judgeval/constants.py

@@ -41,14 +41,16 @@ ROOT_API = os.getenv("JUDGMENT_API_URL", "https://api.judgmentlabs.ai")
 JUDGMENT_EVAL_API_URL = f"{ROOT_API}/evaluate/"
 JUDGMENT_DATASETS_PUSH_API_URL = f"{ROOT_API}/datasets/push/"
 JUDGMENT_DATASETS_PULL_API_URL = f"{ROOT_API}/datasets/pull/"
+JUDGMENT_DATASETS_DELETE_API_URL = f"{ROOT_API}/datasets/delete/"
 JUDGMENT_DATASETS_EXPORT_JSONL_API_URL = f"{ROOT_API}/datasets/export_jsonl/"
-JUDGMENT_DATASETS_PULL_ALL_API_URL = f"{ROOT_API}/datasets/get_all_stats/"
-JUDGMENT_DATASETS_EDIT_API_URL = f"{ROOT_API}/datasets/edit/"
+JUDGMENT_DATASETS_PROJECT_STATS_API_URL = f"{ROOT_API}/datasets/fetch_stats_by_project/"
+JUDGMENT_DATASETS_INSERT_API_URL = f"{ROOT_API}/datasets/insert_examples/"
 JUDGMENT_EVAL_LOG_API_URL = f"{ROOT_API}/log_eval_results/"
 JUDGMENT_EVAL_FETCH_API_URL = f"{ROOT_API}/fetch_eval_results/"
 JUDGMENT_EVAL_DELETE_API_URL = f"{ROOT_API}/delete_eval_results_by_project_and_run_names/"
 JUDGMENT_EVAL_DELETE_PROJECT_API_URL = f"{ROOT_API}/delete_eval_results_by_project/"
 JUDGMENT_PROJECT_DELETE_API_URL = f"{ROOT_API}/projects/delete/"
+JUDGMENT_PROJECT_CREATE_API_URL = f"{ROOT_API}/projects/add/"
 JUDGMENT_TRACES_FETCH_API_URL = f"{ROOT_API}/traces/fetch/"
 JUDGMENT_TRACES_SAVE_API_URL = f"{ROOT_API}/traces/save/"
 JUDGMENT_TRACES_DELETE_API_URL = f"{ROOT_API}/traces/delete/"

judgeval/data/__init__.py

@@ -1,13 +1,10 @@
 from judgeval.data.example import Example, ExampleParams
-from judgeval.data.api_example import ProcessExample, create_process_example
 from judgeval.data.scorer_data import ScorerData, create_scorer_data
 from judgeval.data.result import ScoringResult, generate_scoring_result
 
 __all__ = [
     "Example",
     "ExampleParams",
-    "ProcessExample",
-    "create_process_example",
     "ScorerData",
     "create_scorer_data",
     "ScoringResult",

judgeval/data/{api_example.py → custom_api_example.py}

@@ -2,6 +2,7 @@ from typing import List, Optional, Dict, Any, Union
 from pydantic import BaseModel, ConfigDict, model_validator
 
 from judgeval.data.example import Example
+from judgeval.data.custom_example import CustomExample
 from judgeval.data.scorer_data import ScorerData
 from judgeval.common.logger import debug, error
 
@@ -12,13 +13,13 @@ class ProcessExample(BaseModel):
     internal operations and keeping track of the evaluation process.
     """
     name: str
-    input: Optional[str] = None
-    actual_output: Optional[Union[str, List[str]]] = None
-    expected_output: Optional[Union[str, List[str]]] = None
-    context: Optional[list] = None
-    retrieval_context: Optional[list] = None
-    tools_called: Optional[list] = None
-    expected_tools: Optional[list] = None
+    # input: Optional[str] = None
+    # actual_output: Optional[Union[str, List[str]]] = None
+    # expected_output: Optional[Union[str, List[str]]] = None
+    # context: Optional[list] = None
+    # retrieval_context: Optional[list] = None
+    # tools_called: Optional[list] = None
+    # expected_tools: Optional[list] = None
 
     # make these optional, not all test cases in a conversation will be evaluated
     success: Optional[bool] = None
@@ -57,10 +58,10 @@ class ProcessExample(BaseModel):
 
     def update_run_duration(self, run_duration: float):
         self.run_duration = run_duration
-    
 
-def create_process_example(
-    example: Example,
+
+def create_process_custom_example(
+    example: CustomExample,
 ) -> ProcessExample:
     """
     When an LLM Test Case is executed, we track its progress using an ProcessExample.
@@ -79,13 +80,6 @@ def create_process_example(
     debug(f"Creating ProcessExample for: {name}")
     process_ex = ProcessExample(
         name=name,
-        input=example.input,
-        actual_output=example.actual_output,
-        expected_output=example.expected_output,
-        context=example.context,
-        retrieval_context=example.retrieval_context,
-        tools_called=example.tools_called,
-        expected_tools=example.expected_tools,
         success=success,
         scorers_data=scorers_data,
         run_duration=None,
@@ -94,5 +88,4 @@ def create_process_example(
         additional_metadata=example.additional_metadata,
         trace_id=example.trace_id
     )
-    return process_ex
-
+    return process_ex

judgeval/data/datasets/eval_dataset_client.py

@@ -7,8 +7,9 @@ from judgeval.common.logger import debug, error, warning, info
 from judgeval.constants import (
     JUDGMENT_DATASETS_PUSH_API_URL,
     JUDGMENT_DATASETS_PULL_API_URL, 
-    JUDGMENT_DATASETS_PULL_ALL_API_URL,
-    JUDGMENT_DATASETS_EDIT_API_URL,
+    JUDGMENT_DATASETS_PROJECT_STATS_API_URL,
+    JUDGMENT_DATASETS_DELETE_API_URL,
+    JUDGMENT_DATASETS_INSERT_API_URL,
     JUDGMENT_DATASETS_EXPORT_JSONL_API_URL
 )
 from judgeval.data import Example
@@ -25,7 +26,7 @@ class EvalDatasetClient:
     def create_dataset(self) -> EvalDataset:
         return EvalDataset(judgment_api_key=self.judgment_api_key)
     
-    def push(self, dataset: EvalDataset, alias: str, overwrite: Optional[bool] = False) -> bool:
+    def push(self, dataset: EvalDataset, alias: str, project_name: str, overwrite: Optional[bool] = False) -> bool:
         debug(f"Pushing dataset with alias '{alias}' (overwrite={overwrite})")
         if overwrite:
             warning(f"Overwrite enabled for alias '{alias}'")
@@ -53,7 +54,8 @@ class EvalDatasetClient:
                 total=100,
             )
             content = {
-                    "alias": alias,
+                    "dataset_alias": alias,
+                    "project_name": project_name,
                     "examples": [e.to_dict() for e in dataset.examples],
                     "overwrite": overwrite,
                 }
@@ -88,7 +90,7 @@ class EvalDatasetClient:
                 )
             return True
         
-    def pull(self, alias: str) -> EvalDataset:
+    def pull(self, alias: str, project_name: str) -> EvalDataset:
         debug(f"Pulling dataset with alias '{alias}'")
         """
         Pulls the dataset from Judgment platform
@@ -96,7 +98,7 @@ class EvalDatasetClient:
         Mock request:
         {
             "alias": alias,
-            "user_id": user_id
+            "project_name": project_name
         } 
         ==>
         {
@@ -118,7 +120,8 @@ class EvalDatasetClient:
                     total=100,
                 )
                 request_body = {
-                    "alias": alias,
+                    "dataset_alias": alias,
+                    "project_name": project_name
                 }
 
                 try:
@@ -139,24 +142,58 @@ class EvalDatasetClient:
 
                 info(f"Successfully pulled dataset with alias '{alias}'")
                 payload = response.json()
+
                 dataset.examples = [Example(**e) for e in payload.get("examples", [])]
-                dataset._alias = payload.get("_alias")
-                dataset._id = payload.get("_id")
+                dataset._alias = payload.get("alias")
+                dataset._id = payload.get("id")
                 progress.update(
                     task_id,
                     description=f"{progress.tasks[task_id].description} [rgb(25,227,160)]Done!)",
                 )
 
                 return dataset
+    
+    def delete(self, alias: str, project_name: str) -> bool:
+        with Progress(
+                SpinnerColumn(style="rgb(106,0,255)"),
+                TextColumn("[progress.description]{task.description}"),
+                transient=False,
+            ) as progress:
+                task_id = progress.add_task(
+                    f"Deleting [rgb(106,0,255)]'{alias}'[/rgb(106,0,255)] from Judgment...",
+                    total=100,
+                )
+                request_body = {
+                    "dataset_alias": alias,
+                    "project_name": project_name
+                }
 
-    def pull_all_user_dataset_stats(self) -> dict:
-        debug(f"Pulling user datasets stats for user_id: {self.judgment_api_key}'")
+                try:
+                    response = requests.post(
+                        JUDGMENT_DATASETS_DELETE_API_URL, 
+                        json=request_body,
+                        headers={
+                            "Content-Type": "application/json",
+                            "Authorization": f"Bearer {self.judgment_api_key}",
+                            "X-Organization-Id": self.organization_id
+                        },
+                        verify=True
+                    )
+                    response.raise_for_status()
+                except requests.exceptions.RequestException as e:
+                    error(f"Error deleting dataset: {str(e)}")
+                    raise
+
+                return True
+        
+    def pull_project_dataset_stats(self, project_name: str) -> dict:
+        debug(f"Pulling project datasets stats for project_name: {project_name}'")
         """
-        Pulls the user datasets stats from Judgment platform
+        Pulls the project datasets stats from Judgment platform    
 
         Mock request:
         {
-            "user_id": user_id
+            "project_name": project_name
         } 
         ==>
         {
@@ -177,11 +214,12 @@ class EvalDatasetClient:
                     total=100,
                 )
                 request_body = {
+                    "project_name": project_name
                 }
 
                 try:
                     response = requests.post(
-                        JUDGMENT_DATASETS_PULL_ALL_API_URL, 
+                        JUDGMENT_DATASETS_PROJECT_STATS_API_URL, 
                         json=request_body,
                         headers={
                             "Content-Type": "application/json",
@@ -205,7 +243,7 @@ class EvalDatasetClient:
 
                 return payload
 
-    def edit_dataset(self, alias: str, examples: List[Example]) -> bool:
+    def insert_dataset(self, alias: str, examples: List[Example], project_name: str) -> bool:
         """
         Edits the dataset on Judgment platform by adding new examples
 
@@ -213,7 +251,7 @@ class EvalDatasetClient:
         {
             "alias": alias,
             "examples": [...],
-            "judgment_api_key": self.judgment_api_key
+            "project_name": project_name
         }
         """
         with Progress(
@@ -227,13 +265,14 @@ class EvalDatasetClient:
             )
             
             content = {
-                "alias": alias,
+                "dataset_alias": alias,
                 "examples": [e.to_dict() for e in examples],
+                "project_name": project_name
             }
 
             try:
                 response = requests.post(
-                    JUDGMENT_DATASETS_EDIT_API_URL,
+                    JUDGMENT_DATASETS_INSERT_API_URL,
                     json=content,
                     headers={
                         "Content-Type": "application/json",
@@ -250,7 +289,7 @@ class EvalDatasetClient:
             info(f"Successfully edited dataset '{alias}'")
             return True
 
-    def export_jsonl(self, alias: str) -> requests.Response:
+    def export_jsonl(self, alias: str, project_name: str) -> requests.Response:
         """Export dataset in JSONL format from Judgment platform"""
         debug(f"Exporting dataset with alias '{alias}' as JSONL")
         with Progress(
@@ -265,7 +304,7 @@ class EvalDatasetClient:
             try:
                 response = requests.post(
                     JUDGMENT_DATASETS_EXPORT_JSONL_API_URL,
-                    json={"alias": alias},
+                    json={"dataset_alias": alias, "project_name": project_name},
                     headers={
                         "Content-Type": "application/json",
                         "Authorization": f"Bearer {self.judgment_api_key}",

judgeval/data/result.py

@@ -1,10 +1,11 @@
 from dataclasses import dataclass
 from typing import List, Union, Optional, Dict, Any, Union
+from judgeval.common.logger import debug, error
+from pydantic import BaseModel
+from judgeval.data import ScorerData, Example
 
-from judgeval.data import ScorerData, ProcessExample
 
-@dataclass
-class ScoringResult:
+class ScoringResult(BaseModel):
     """
     A ScoringResult contains the output of one or more scorers applied to a single example.
     Ie: One input, one actual_output, one expected_output, etc..., and 1+ scorer (Faithfulness, Hallucination, Summarization, etc...)
@@ -13,69 +14,44 @@ class ScoringResult:
         success (bool): Whether the evaluation was successful. 
                         This means that all scorers applied to this example returned a success.
         scorer_data (List[ScorerData]): The scorers data for the evaluated example
-        input (Optional[str]): The input to the example
-        actual_output (Optional[str]): The actual output of the example
-        expected_output (Optional[str]): The expected output of the example
-        context (Optional[List[str]]): The context of the example
-        retrieval_context (Optional[List[str]]): The retrieval context of the example
-        additional_metadata (Optional[Dict[str, Any]]): The additional metadata of the example
-        tools_called (Optional[List[str]]): The tools called by the example
-        expected_tools (Optional[List[str]]): The expected tools of the example
-        trace_id (Optional[str]): The trace id of the example
+        data_object (Optional[Example]): The original example object that was used to create the ScoringResult, can be Example, CustomExample (future), WorkflowRun (future)
         
     """
     # Fields for scoring outputs 
     success: bool  # used for unit testing
     scorers_data: Union[List[ScorerData], None]
+    name: Optional[str] = None
 
-    # Inputs from the original example
-    input: Optional[str] = None
-    actual_output: Optional[Union[str, List[str]]] = None
-    expected_output: Optional[Union[str, List[str]]] = None
-    context: Optional[List[str]] = None
-    retrieval_context: Optional[List[str]] = None
-    additional_metadata: Optional[Dict[str, Any]] = None
-    tools_called: Optional[List[str]] = None
-    expected_tools: Optional[List[str]] = None
+    # The original example object that was used to create the ScoringResult
+    data_object: Optional[Example] = None #can be Example, CustomExample (future), WorkflowRun (future)
     trace_id: Optional[str] = None
     
-    example_id: Optional[str] = None
-    eval_run_name: Optional[str] = None
+    # Additional fields for internal use
+    run_duration: Optional[float] = None
+    evaluation_cost: Optional[float] = None
     
     def to_dict(self) -> dict:
         """Convert the ScoringResult instance to a dictionary, properly serializing scorer_data."""
         return {
             "success": self.success,
             "scorers_data": [scorer_data.to_dict() for scorer_data in self.scorers_data] if self.scorers_data else None,
-            "input": self.input,
-            "actual_output": self.actual_output,
-            "expected_output": self.expected_output,
-            "context": self.context,
-            "retrieval_context": self.retrieval_context,
-            "additional_metadata": self.additional_metadata,
-            "tools_called": self.tools_called,
-            "expected_tools": self.expected_tools,
-            "trace_id": self.trace_id,
-            "example_id": self.example_id
+            "data_object": self.data_object.to_dict() if self.data_object else None,
         }
-    
+
     def __str__(self) -> str:
         return f"ScoringResult(\
             success={self.success}, \
             scorer_data={self.scorers_data}, \
-            input={self.input}, \
-            actual_output={self.actual_output}, \
-            expected_output={self.expected_output}, \
-            context={self.context}, \
-            retrieval_context={self.retrieval_context}, \
-            additional_metadata={self.additional_metadata}, \
-            tools_called={self.tools_called}, \
-            expected_tools={self.expected_tools}, \
-            trace_id={self.trace_id})"
+            data_object={self.data_object}, \
+            run_duration={self.run_duration}, \
+            evaluation_cost={self.evaluation_cost})"
 
 
 def generate_scoring_result(
-    process_example: ProcessExample,
+    example: Example,
+    scorers_data: List[ScorerData],
+    run_duration: float,
+    success: bool,
 ) -> ScoringResult:
     """
     Creates a final ScoringResult object for an evaluation run based on the results from a completed LLMApiTestCase.
@@ -83,16 +59,18 @@ def generate_scoring_result(
     When an LLMTestCase is executed, it turns into an LLMApiTestCase and the progress of the evaluation run is tracked.
     At the end of the evaluation run, we create a TestResult object out of the completed LLMApiTestCase.
     """
-    return ScoringResult(
-        success=process_example.success,
-        scorers_data=process_example.scorers_data,
-        input=process_example.input,
-        actual_output=process_example.actual_output,
-        expected_output=process_example.expected_output,
-        context=process_example.context,
-        retrieval_context=process_example.retrieval_context,
-        additional_metadata=process_example.additional_metadata,
-        tools_called=process_example.tools_called,
-        expected_tools=process_example.expected_tools,
-        trace_id=process_example.trace_id
+    if example.name is not None:
+        name = example.name
+    else:
+        name = "Test Case Placeholder"
+        debug(f"No name provided for example, using default name: {name}")
+    debug(f"Creating ScoringResult for: {name}")
+    scoring_result = ScoringResult(
+        name=name,
+        data_object=example,
+        success=success,
+        scorers_data=scorers_data,
+        run_duration=run_duration,
+        evaluation_cost=None,
     )
+    return scoring_result

judgeval/evaluation_run.py

@@ -34,6 +34,7 @@ class EvaluationRun(BaseModel):
     model: Union[str, List[str], JudgevalJudge]
     aggregator: Optional[str] = None
     metadata: Optional[Dict[str, Any]] = None
+    trace_span_id: Optional[str] = None
     # API Key will be "" until user calls client.run_eval(), then API Key will be set
     judgment_api_key: Optional[str] = ""
     override: Optional[bool] = False

judgeval/judgment_client.py

@@ -27,7 +27,8 @@ from judgeval.constants import (
     JUDGMENT_EVAL_FETCH_API_URL, 
     JUDGMENT_EVAL_DELETE_API_URL, 
     JUDGMENT_EVAL_DELETE_PROJECT_API_URL,
-    JUDGMENT_PROJECT_DELETE_API_URL
+    JUDGMENT_PROJECT_DELETE_API_URL,
+    JUDGMENT_PROJECT_CREATE_API_URL
 )
 from judgeval.common.exceptions import JudgmentAPIError
 from pydantic import BaseModel
@@ -43,8 +44,16 @@ class DeleteEvalRunRequestBody(BaseModel):
     project_name: str
     judgment_api_key: str
 
+class SingletonMeta(type):
+    _instances = {}
 
-class JudgmentClient:
+    def __call__(cls, *args, **kwargs):
+        if cls not in cls._instances:
+            instance = super().__call__(*args, **kwargs)
+            cls._instances[cls] = instance
+        return cls._instances[cls]
+
+class JudgmentClient(metaclass=SingletonMeta):
     def __init__(self, judgment_api_key: str = os.getenv("JUDGMENT_API_KEY"), organization_id: str = os.getenv("JUDGMENT_ORG_ID")):
         self.judgment_api_key = judgment_api_key
         self.organization_id = organization_id
@@ -56,8 +65,8 @@ class JudgmentClient:
             # May be bad to output their invalid API key...
             raise JudgmentAPIError(f"Issue with passed in Judgment API key: {response}")
         else:
-            print(f"Successfully initialized JudgmentClient, welcome back {response.get('detail', {}).get('user_name', 'user')}!")
-    
+            print(f"Successfully initialized JudgmentClient!")
+
     def a_run_evaluation(
         self, 
         examples: List[Example],
@@ -267,7 +276,7 @@ class JudgmentClient:
     def create_dataset(self) -> EvalDataset:
         return self.eval_dataset_client.create_dataset()
 
-    def push_dataset(self, alias: str, dataset: EvalDataset, overwrite: Optional[bool] = False) -> bool:
+    def push_dataset(self, alias: str, dataset: EvalDataset, project_name: str, overwrite: Optional[bool] = False) -> bool:
         """
         Uploads an `EvalDataset` to the Judgment platform for storage.
 
@@ -281,9 +290,9 @@ class JudgmentClient:
         """
         # Set judgment_api_key just in case it was not set
         dataset.judgment_api_key = self.judgment_api_key
-        return self.eval_dataset_client.push(dataset, alias, overwrite)
+        return self.eval_dataset_client.push(dataset, alias, project_name, overwrite)
     
-    def pull_dataset(self, alias: str) -> EvalDataset:
+    def pull_dataset(self, alias: str, project_name: str) -> EvalDataset:
         """
         Retrieves a saved `EvalDataset` from the Judgment platform.
 
@@ -293,25 +302,31 @@ class JudgmentClient:
         Returns:
             EvalDataset: The retrieved dataset
         """
-        return self.eval_dataset_client.pull(alias)
+        return self.eval_dataset_client.pull(alias, project_name)
+
+    def delete_dataset(self, alias: str, project_name: str) -> bool:
+        """
+        Deletes a saved `EvalDataset` from the Judgment platform.
+        """
+        return self.eval_dataset_client.delete(alias, project_name)
     
-    def pull_all_user_dataset_stats(self) -> dict:
+    def pull_project_dataset_stats(self, project_name: str) -> dict:
         """
-        Retrieves all dataset stats from the Judgment platform for the user.
+        Retrieves all dataset stats from the Judgment platform for the project.
 
         Args:
-            alias (str): The name of the dataset to retrieve
+            project_name (str): The name of the project to retrieve
 
         Returns:
-            EvalDataset: The retrieved dataset
+            dict: The retrieved dataset stats
         """
-        return self.eval_dataset_client.pull_all_user_dataset_stats()
+        return self.eval_dataset_client.pull_project_dataset_stats(project_name)
     
-    def edit_dataset(self, alias: str, examples: List[Example]) -> bool:
+    def insert_dataset(self, alias: str, examples: List[Example], project_name: str) -> bool:
         """
         Edits the dataset on Judgment platform by adding new examples
         """
-        return self.eval_dataset_client.edit_dataset(alias, examples)
+        return self.eval_dataset_client.insert_dataset(alias, examples, project_name)
     
     # Maybe add option where you can pass in the EvaluationRun object and it will pull the eval results from the backend
     def pull_eval(self, project_name: str, eval_run_name: str) -> List[Dict[str, Union[str, List[ScoringResult]]]]:
@@ -402,6 +417,23 @@ class JudgmentClient:
             raise ValueError(f"Error deleting eval results: {response.json()}")
         return response.json()
     
+    def create_project(self, project_name: str) -> bool:
+        """
+        Creates a project on the server.
+        """
+        response = requests.post(JUDGMENT_PROJECT_CREATE_API_URL, 
+                        json={
+                            "project_name": project_name,
+                        },
+                        headers={
+                            "Content-Type": "application/json",
+                            "Authorization": f"Bearer {self.judgment_api_key}",
+                            "X-Organization-Id": self.organization_id
+                        })
+        if response.status_code != requests.codes.ok:
+            raise ValueError(f"Error creating project: {response.json()}")
+        return response.json()
+    
     def delete_project(self, project_name: str) -> bool:
         """
         Deletes a project from the server. Which also deletes all evaluations and traces associated with the project.

judgeval/run_evaluation.py

@@ -117,21 +117,23 @@ def merge_results(api_results: List[ScoringResult], local_results: List[ScoringR
     
     # Each ScoringResult in api and local have all the same fields besides `scorers_data`
     for api_result, local_result in zip(api_results, local_results):
-        if api_result.input != local_result.input:
+        if not (api_result.data_object and local_result.data_object):
+            raise ValueError("Data object is None in one of the results.")
+        if api_result.data_object.input != local_result.data_object.input:
             raise ValueError("The API and local results are not aligned.")
-        if api_result.actual_output != local_result.actual_output:
+        if api_result.data_object.actual_output != local_result.data_object.actual_output:
             raise ValueError("The API and local results are not aligned.")
-        if api_result.expected_output != local_result.expected_output:
+        if api_result.data_object.expected_output != local_result.data_object.expected_output:
             raise ValueError("The API and local results are not aligned.")
-        if api_result.context != local_result.context:
+        if api_result.data_object.context != local_result.data_object.context:
             raise ValueError("The API and local results are not aligned.")
-        if api_result.retrieval_context != local_result.retrieval_context:
+        if api_result.data_object.retrieval_context != local_result.data_object.retrieval_context:
             raise ValueError("The API and local results are not aligned.")
-        if api_result.additional_metadata != local_result.additional_metadata:
+        if api_result.data_object.additional_metadata != local_result.data_object.additional_metadata:
             raise ValueError("The API and local results are not aligned.")
-        if api_result.tools_called != local_result.tools_called:
+        if api_result.data_object.tools_called != local_result.data_object.tools_called:
             raise ValueError("The API and local results are not aligned.")
-        if api_result.expected_tools != local_result.expected_tools:
+        if api_result.data_object.expected_tools != local_result.data_object.expected_tools:
             raise ValueError("The API and local results are not aligned.")
         
         
@@ -422,23 +424,7 @@ def run_eval(evaluation_run: EvaluationRun, override: bool = False, ignore_error
             
             # Convert the response data to `ScoringResult` objects
             debug("Processing API results")
-            for idx, result in enumerate(response_data["results"]):  
-                with example_logging_context(evaluation_run.examples[idx].timestamp, evaluation_run.examples[idx].example_id):
-                    for scorer in judgment_scorers:
-                        debug(f"Processing API result for example {idx} and scorer {scorer.score_type}")
-                    # filter for key-value pairs that are used to initialize ScoringResult
-                    # there may be some stuff in here that doesn't belong in ScoringResult
-                    # TODO: come back and refactor this to have ScoringResult take in **kwargs
-                    filtered_result = {k: v for k, v in result.items() if k in ScoringResult.__annotations__}
-                    
-                    # Convert scorers_data dicts to ScorerData objects
-                    if "scorers_data" in filtered_result and filtered_result["scorers_data"]:
-                        filtered_result["scorers_data"] = [
-                            ScorerData(**scorer_dict) 
-                            for scorer_dict in filtered_result["scorers_data"]
-                        ]
-                    
-                    api_results.append(ScoringResult(**filtered_result))
+            api_results = [ScoringResult(**result) for result in response_data["results"]]
         # Run local evals
         if local_scorers:  # List[JudgevalScorer]
             # We should be removing local scorers soon
@@ -477,7 +463,7 @@ def run_eval(evaluation_run: EvaluationRun, override: bool = False, ignore_error
         #         judgment_api_key=evaluation_run.judgment_api_key,
         #         organization_id=evaluation_run.organization_id
         #     )
-        
+        # print(merged_results)
         if evaluation_run.log_results:
             pretty_str = run_with_spinner("Logging Results: ", log_evaluation_results, merged_results, evaluation_run)
             rprint(pretty_str)
@@ -504,15 +490,14 @@ def assert_test(scoring_results: List[ScoringResult]) -> None:
 
             # Create a test case context with all relevant fields
             test_case = {
-                'input': result.input,
-                'actual_output': result.actual_output,
-                'expected_output': result.expected_output,
-                'context': result.context,
-                'retrieval_context': result.retrieval_context,
-                'additional_metadata': result.additional_metadata,
-                'tools_called': result.tools_called,
-                'expected_tools': result.expected_tools,
-                'eval_run_name': result.eval_run_name,
+                'input': result.data_object.input,
+                'actual_output': result.data_object.actual_output,
+                'expected_output': result.data_object.expected_output,
+                'context': result.data_object.context,
+                'retrieval_context': result.data_object.retrieval_context,
+                'additional_metadata': result.data_object.additional_metadata,
+                'tools_called': result.data_object.tools_called,
+                'expected_tools': result.data_object.expected_tools,
                 'failed_scorers': []
             }
             if result.scorers_data:
@@ -533,7 +518,6 @@ def assert_test(scoring_results: List[ScoringResult]) -> None:
             error_msg += f"Additional Metadata: {fail_case['additional_metadata']}\n"
             error_msg += f"Tools Called: {fail_case['tools_called']}\n"
             error_msg += f"Expected Tools: {fail_case['expected_tools']}\n"
-            error_msg += f"Eval Run Name: {fail_case['eval_run_name']}\n"
     
             for fail_scorer in fail_case['failed_scorers']: