judgeval 0.5.0__py3-none-any.whl → 0.6.0__py3-none-any.whl

judgeval/cli.py

@@ -0,0 +1,65 @@
+#!/usr/bin/env python3
+
+import typer
+from pathlib import Path
+from dotenv import load_dotenv
+from judgeval.common.logger import judgeval_logger
+from judgeval.judgment_client import JudgmentClient
+
+load_dotenv()
+
+app = typer.Typer(
+    no_args_is_help=True,
+    rich_markup_mode=None,
+    rich_help_panel=None,
+    pretty_exceptions_enable=False,
+    pretty_exceptions_show_locals=False,
+    pretty_exceptions_short=False,
+)
+
+
+@app.command("upload_scorer")
+def upload_scorer(
+    scorer_file_path: str,
+    requirements_file_path: str,
+    unique_name: str = typer.Option(
+        None, help="Custom name for the scorer (auto-detected if not provided)"
+    ),
+):
+    # Validate file paths
+    if not Path(scorer_file_path).exists():
+        judgeval_logger.error(f"Scorer file not found: {scorer_file_path}")
+        raise typer.Exit(1)
+
+    if not Path(requirements_file_path).exists():
+        judgeval_logger.error(f"Requirements file not found: {requirements_file_path}")
+        raise typer.Exit(1)
+
+    try:
+        client = JudgmentClient()
+
+        result = client.save_custom_scorer(
+            scorer_file_path=scorer_file_path,
+            requirements_file_path=requirements_file_path,
+            unique_name=unique_name,
+        )
+
+        if not result:
+            judgeval_logger.error("Failed to upload custom scorer")
+            raise typer.Exit(1)
+
+        raise typer.Exit(0)
+    except Exception:
+        raise
+
+
+@app.command()
+def version():
+    """Show version info"""
+    judgeval_logger.info("JudgEval CLI v0.0.0")
+
+
+if __name__ == "__main__":
+    app()
+
+# judgeval upload_scorer /Users/alanzhang/repo/JudgmentLabs/judgeval/src/demo/profile_match_scorer.py /Users/alanzhang/repo/JudgmentLabs/judgeval/src/demo/requirements.txt

judgeval/common/api/api.py

@@ -20,13 +20,11 @@ from judgeval.common.api.constants import (
     JUDGMENT_EVAL_DELETE_API_URL,
     JUDGMENT_ADD_TO_RUN_EVAL_QUEUE_API_URL,
     JUDGMENT_GET_EVAL_STATUS_API_URL,
-    JUDGMENT_CHECK_EXPERIMENT_TYPE_API_URL,
-    JUDGMENT_EVAL_RUN_NAME_EXISTS_API_URL,
     JUDGMENT_SCORER_SAVE_API_URL,
     JUDGMENT_SCORER_FETCH_API_URL,
     JUDGMENT_SCORER_EXISTS_API_URL,
+    JUDGMENT_CUSTOM_SCORER_UPLOAD_API_URL,
     JUDGMENT_DATASETS_APPEND_TRACES_API_URL,
-    JUDGMENT_CHECK_EXAMPLE_KEYS_API_URL,
 )
 from judgeval.common.api.constants import (
     TraceFetchPayload,
@@ -45,12 +43,11 @@ from judgeval.common.api.constants import (
     DeleteEvalRunRequestBody,
     EvalLogPayload,
     EvalStatusPayload,
-    CheckExperimentTypePayload,
-    EvalRunNameExistsPayload,
     ScorerSavePayload,
     ScorerFetchPayload,
     ScorerExistsPayload,
-    CheckExampleKeysPayload,
+    CustomScorerUploadPayload,
+    CustomScorerTemplateResponse,
 )
 from judgeval.utils.requests import requests
 from judgeval.common.api.json_encoder import json_encoder
@@ -97,14 +94,20 @@ class JudgmentApiClient:
         method: Literal["POST", "PATCH", "GET", "DELETE"],
         url: str,
         payload: Any,
+        timeout: Optional[Union[float, tuple]] = None,
     ) -> Any:
+        # Prepare request kwargs with optional timeout
+        request_kwargs = self._request_kwargs()
+        if timeout is not None:
+            request_kwargs["timeout"] = timeout
+
         if method == "GET":
             r = requests.request(
                 method,
                 url,
                 params=payload,
                 headers=self._headers(),
-                **self._request_kwargs(),
+                **request_kwargs,
             )
         else:
             r = requests.request(
@@ -112,7 +115,7 @@ class JudgmentApiClient:
                 url,
                 json=json_encoder(payload),
                 headers=self._headers(),
-                **self._request_kwargs(),
+                **request_kwargs,
             )
 
         try:
@@ -186,10 +189,10 @@ class JudgmentApiClient:
         payload: EvalLogPayload = {"results": results, "run": run}
         return self._do_request("POST", JUDGMENT_EVAL_LOG_API_URL, payload)
 
-    def fetch_evaluation_results(self, project_name: str, eval_name: str):
+    def fetch_evaluation_results(self, experiment_run_id: str, project_name: str):
         payload: EvalRunRequestBody = {
             "project_name": project_name,
-            "eval_name": eval_name,
+            "experiment_run_id": experiment_run_id,
         }
         return self._do_request("POST", JUDGMENT_EVAL_FETCH_API_URL, payload)
 
@@ -204,43 +207,21 @@ class JudgmentApiClient:
     def add_to_evaluation_queue(self, payload: Dict[str, Any]):
         return self._do_request("POST", JUDGMENT_ADD_TO_RUN_EVAL_QUEUE_API_URL, payload)
 
-    def get_evaluation_status(self, eval_name: str, project_name: str):
+    def get_evaluation_status(self, experiment_run_id: str, project_name: str):
         payload: EvalStatusPayload = {
-            "eval_name": eval_name,
+            "experiment_run_id": experiment_run_id,
             "project_name": project_name,
             "judgment_api_key": self.api_key,
         }
         return self._do_request("GET", JUDGMENT_GET_EVAL_STATUS_API_URL, payload)
 
-    def check_experiment_type(self, eval_name: str, project_name: str, is_trace: bool):
-        payload: CheckExperimentTypePayload = {
-            "eval_name": eval_name,
-            "project_name": project_name,
-            "judgment_api_key": self.api_key,
-            "is_trace": is_trace,
-        }
-        return self._do_request("POST", JUDGMENT_CHECK_EXPERIMENT_TYPE_API_URL, payload)
-
-    def check_eval_run_name_exists(self, eval_name: str, project_name: str):
-        payload: EvalRunNameExistsPayload = {
-            "eval_name": eval_name,
-            "project_name": project_name,
-            "judgment_api_key": self.api_key,
-        }
-        return self._do_request("POST", JUDGMENT_EVAL_RUN_NAME_EXISTS_API_URL, payload)
-
-    def check_example_keys(self, keys: List[str], eval_name: str, project_name: str):
-        payload: CheckExampleKeysPayload = {
-            "keys": keys,
-            "eval_name": eval_name,
-            "project_name": project_name,
-        }
-        return self._do_request("POST", JUDGMENT_CHECK_EXAMPLE_KEYS_API_URL, payload)
-
-    def save_scorer(self, name: str, prompt: str, options: Optional[dict] = None):
+    def save_scorer(
+        self, name: str, prompt: str, threshold: float, options: Optional[dict] = None
+    ):
         payload: ScorerSavePayload = {
             "name": name,
             "prompt": prompt,
+            "threshold": threshold,
             "options": options,
         }
         try:
@@ -292,6 +273,31 @@ class JudgmentApiClient:
                 request=e.request,
             )
 
+    def upload_custom_scorer(
+        self,
+        scorer_name: str,
+        scorer_code: str,
+        requirements_text: str,
+    ) -> CustomScorerTemplateResponse:
+        """Upload custom scorer to backend"""
+        payload: CustomScorerUploadPayload = {
+            "scorer_name": scorer_name,
+            "scorer_code": scorer_code,
+            "requirements_text": requirements_text,
+        }
+
+        try:
+            # Use longer timeout for custom scorer upload (5 minutes)
+            response = self._do_request(
+                "POST",
+                JUDGMENT_CUSTOM_SCORER_UPLOAD_API_URL,
+                payload,
+                timeout=(10, 300),
+            )
+            return response
+        except JudgmentAPIException as e:
+            raise e
+
     def push_dataset(
         self,
         dataset_alias: str,

judgeval/common/api/constants.py

@@ -49,9 +49,9 @@ JUDGMENT_EVAL_DELETE_API_URL = (
 JUDGMENT_EVAL_DELETE_PROJECT_API_URL = f"{ROOT_API}/delete_eval_results_by_project/"
 JUDGMENT_ADD_TO_RUN_EVAL_QUEUE_API_URL = f"{ROOT_API}/add_to_run_eval_queue/"
 JUDGMENT_GET_EVAL_STATUS_API_URL = f"{ROOT_API}/get_evaluation_status/"
-JUDGMENT_CHECK_EXPERIMENT_TYPE_API_URL = f"{ROOT_API}/check_experiment_type/"
-JUDGMENT_EVAL_RUN_NAME_EXISTS_API_URL = f"{ROOT_API}/eval-run-name-exists/"
-JUDGMENT_CHECK_EXAMPLE_KEYS_API_URL = f"{ROOT_API}/check_example_keys/"
+
+# Custom Scorers API
+JUDGMENT_CUSTOM_SCORER_UPLOAD_API_URL = f"{ROOT_API}/build_sandbox_template/"
 
 
 # Evaluation API Payloads
@@ -73,9 +73,9 @@ class EvalLogPayload(TypedDict):
 
 
 class EvalStatusPayload(TypedDict):
-    eval_name: str
-    project_name: str
+    experiment_run_id: str
     judgment_api_key: str
+    project_name: str
 
 
 class CheckExperimentTypePayload(TypedDict):
@@ -162,6 +162,7 @@ JUDGMENT_SCORER_EXISTS_API_URL = f"{ROOT_API}/scorer_exists/"
 class ScorerSavePayload(TypedDict):
     name: str
     prompt: str
+    threshold: float
     options: Optional[dict]
 
 
@@ -171,3 +172,15 @@ class ScorerFetchPayload(TypedDict):
 
 class ScorerExistsPayload(TypedDict):
     name: str
+
+
+class CustomScorerUploadPayload(TypedDict):
+    scorer_name: str
+    scorer_code: str
+    requirements_text: str
+
+
+class CustomScorerTemplateResponse(TypedDict):
+    scorer_name: str
+    status: str
+    message: str

judgeval/common/api/json_encoder.py

@@ -84,7 +84,7 @@ def json_encoder(
         )
 
     # Sequences
-    if isinstance(obj, (list, set, frozenset, GeneratorType, tuple, deque)):
+    if isinstance(obj, (list, set, frozenset, tuple, deque)):
         return _dump_sequence(
             obj=obj,
         )
@@ -169,16 +169,15 @@ def _dump_other(
     obj: Any,
 ) -> Any:
     """
-    Dump an object to a hashable object, using the same parameters as jsonable_encoder
+    Dump an object to a representation without iterating it.
+
+    Avoids calling dict(obj) which can consume iterators/generators or
+    invoke user-defined iteration protocols.
     """
     try:
-        data = dict(obj)
-    except Exception:
         return repr(obj)
-
-    return json_encoder(
-        data,
-    )
+    except Exception:
+        return str(obj)
 
 
 def iso_format(o: Union[datetime.date, datetime.time]) -> str:
@@ -218,7 +217,7 @@ ENCODERS_BY_TYPE: Dict[Type[Any], Callable[[Any], Any]] = {
     Enum: lambda o: o.value,
     frozenset: list,
     deque: list,
-    GeneratorType: list,
+    GeneratorType: repr,
     Path: str,
     Pattern: lambda o: o.pattern,
     SecretBytes: str,