lazy import sklearn

cleanlab_studio/studio/trustworthy_language_model.py

@@ -782,6 +782,11 @@ class TLMOptions(TypedDict):
 
  log (List[str], default = []): optionally specify additional logs or metadata to return.
  For instance, include "explanation" here to get explanations of why a response is scored with low trustworthiness.
+
+ custom_eval_criteria (List[Dict[str, Any]], default = []): optionally specify custom evalution criteria.
+ The expected input format is a list of dictionaries, where each dictionary has the following keys:
+ - name: name of the evaluation criteria
+ - criteria: the instruction for the evaluation criteria
  """
 
  model: NotRequired[str]

cleanlab_studio/utils/tlm_calibrated.py

@@ -13,9 +13,6 @@
 import numpy as np
 import numpy.typing as npt
 import pandas as pd
-from sklearn.ensemble import RandomForestRegressor
-from sklearn.exceptions import NotFittedError
-from sklearn.utils.validation import check_is_fitted
 
 from cleanlab_studio.errors import ValidationError, TlmNotCalibratedError
 from cleanlab_studio.internal.types import TLMQualityPreset
@@ -36,6 +33,14 @@ def __init__(
  Use `Studio.TLMCalibrated()` instead of this method to initialize a TLMCalibrated object.
  lazydocs: ignore
  """
+ try:
+ from sklearn.ensemble import RandomForestRegressor
+ except ImportError:
+ raise ImportError(
+ "Cannot import scikit-learn which is required to use TLMCalibrated. "
+ "Please install it using `pip install scikit-learn` and try again."
+ )
+
  self._api_key = api_key
 
  if quality_preset not in {"base", "low", "medium"}:
@@ -108,6 +113,15 @@ def get_trustworthiness_score(
  Similar to [`TLM.get_trustworthiness_score()`](../trustworthy_language_model/#method-get_trustworthiness_score),
  view documentation there for expected input arguments and outputs.
  """
+ try:
+ from sklearn.exceptions import NotFittedError
+ from sklearn.utils.validation import check_is_fitted
+ except ImportError:
+ raise ImportError(
+ "Cannot import scikit-learn which is required to use TLMCalibrated. "
+ "Please install it using `pip install scikit-learn` and try again."
+ )
+
  try:
  check_is_fitted(self._rf_model)
  except NotFittedError:

setup.py

@@ -63,7 +63,6 @@
  "openpyxl>=3.0.0,!=3.1.0",
  "validators>=0.20.0",
  "matplotlib>=3.4.0",
- "scikit-learn",
  ],
  entry_points="""
  [console_scripts]