fix(evals): auto generate evals

packages/traceloop-sdk/.flake8

@@ -9,4 +9,6 @@ exclude =
 	.venv,
 	.pytest_cache
 max-line-length = 120
-per-file-ignores = __init__.py:F401
+per-file-ignores =
+	__init__.py:F401
+	traceloop/sdk/evaluators_generated/*.py:E501

packages/traceloop-sdk/pyproject.toml

@@ -83,6 +83,7 @@ mypy = "^1.18.2"
 types-requests = "^2.31.0"
 types-colorama = "^0.4.15"
 pandas-stubs = "*"
+datamodel-code-generator = "^0.26.0"
 
 [tool.poetry.group.test.dependencies]
 openai = "^1.31.1"

packages/traceloop-sdk/traceloop/sdk/evaluator/__init__.py

@@ -1,9 +1,10 @@
 from .evaluator import Evaluator
 from .config import EvaluatorDetails
-from .evaluators_made_by_traceloop import EvaluatorMadeByTraceloop
+from .evaluators_made_by_traceloop import EvaluatorMadeByTraceloop, create_evaluator
 
 __all__ = [
     "Evaluator",
     "EvaluatorDetails",
     "EvaluatorMadeByTraceloop",
+    "create_evaluator",
 ]

packages/traceloop-sdk/traceloop/sdk/evaluator/evaluator.py

@@ -1,5 +1,6 @@
 import httpx
 from typing import Dict, Optional, Any, List
+from pydantic import ValidationError
 from .field_mapping import normalize_task_output, get_field_suggestions, format_field_help
 
 from .model import (
@@ -11,6 +12,25 @@
 )
 from .stream_client import SSEClient
 from .config import EvaluatorDetails
+from ..evaluators_generated import get_request_model
+
+
+def _validate_evaluator_input(slug: str, input: Dict[str, str]) -> None:
+    """Validate input against the evaluator's request model if available.
+
+    Args:
+        slug: The evaluator slug (e.g., "pii-detector")
+        input: Dictionary of input field names to values
+
+    Raises:
+        ValueError: If input fails validation against the request model
+    """
+    request_model = get_request_model(slug)
+    if request_model:
+        try:
+            request_model(**input)
+        except ValidationError as e:
+            raise ValueError(f"Invalid input for '{slug}': {e}") from e
 
 
 class Evaluator:
@@ -94,6 +114,8 @@ async def run_experiment_evaluator(
         Returns:
             ExecutionResponse: The evaluation result from SSE stream
         """
+        _validate_evaluator_input(evaluator_slug, input)
+
         request = self._build_evaluator_request(
             task_id, experiment_id, experiment_run_id, input, evaluator_version, evaluator_config
         )
@@ -136,6 +158,8 @@ async def trigger_experiment_evaluator(
         Returns:
             str: The execution_id that can be used to check results later
         """
+        _validate_evaluator_input(evaluator_slug, input)
+
         request = self._build_evaluator_request(
             task_id, experiment_id, experiment_run_id, input, evaluator_version, evaluator_config
         )