kiln-ai 0.17.0__py3-none-any.whl → 0.19.0__py3-none-any.whl

kiln_ai/datamodel/test_eval_model.py

@@ -1,7 +1,6 @@
 import pytest
 from pydantic import ValidationError
 
-from kiln_ai.datamodel import BasePrompt
 from kiln_ai.datamodel.basemodel import KilnParentModel
 from kiln_ai.datamodel.eval import (
     Eval,
@@ -9,11 +8,10 @@ from kiln_ai.datamodel.eval import (
     EvalConfigType,
     EvalOutputScore,
     EvalRun,
+    EvalTemplateId,
 )
 from kiln_ai.datamodel.task import Task
-from kiln_ai.datamodel.task_output import (
-    TaskOutputRatingType,
-)
+from kiln_ai.datamodel.task_output import TaskOutputRatingType
 
 
 @pytest.fixture
@@ -519,13 +517,13 @@ def test_eval_run_score_keys_must_match(valid_eval_config, valid_eval_run_data):
     valid_eval_config.parent = eval
 
     # Correct
-    run = EvalRun(
+    EvalRun(
         parent=valid_eval_config,
         **{**valid_eval_run_data, "scores": {"accuracy": 4.5, "critical": 1.0}},
     )
 
     # Correct but wrong order still okay
-    run = EvalRun(
+    EvalRun(
         parent=valid_eval_config,
         **{**valid_eval_run_data, "scores": {"critical": 1.0, "accuracy": 4.5}},
     )
@@ -535,7 +533,7 @@ def test_eval_run_score_keys_must_match(valid_eval_config, valid_eval_run_data):
         ValueError,
         match="The scores produced by the evaluator must match the scores expected by the eval",
     ):
-        run = EvalRun(
+        EvalRun(
             parent=valid_eval_config,
             **{**valid_eval_run_data, "scores": {"accuracy": 4.5}},
         )
@@ -545,7 +543,7 @@ def test_eval_run_score_keys_must_match(valid_eval_config, valid_eval_run_data):
         ValueError,
         match="The scores produced by the evaluator must match the scores expected by the eval",
     ):
-        run = EvalRun(
+        EvalRun(
             parent=valid_eval_config,
             **{
                 **valid_eval_run_data,
@@ -558,7 +556,7 @@ def test_eval_run_score_keys_must_match(valid_eval_config, valid_eval_run_data):
         ValueError,
         match="The scores produced by the evaluator must match the scores expected by the eval",
     ):
-        run = EvalRun(
+        EvalRun(
             parent=valid_eval_config,
             **{**valid_eval_run_data, "scores": {"accuracy": 4.5, "wrong": 1.0}},
         )
@@ -568,7 +566,7 @@ def test_eval_run_custom_scores_not_allowed(valid_eval_config, valid_eval_run_da
     with pytest.raises(
         ValueError, match="Custom scores are not supported in evaluators"
     ):
-        eval = Eval(
+        Eval(
             name="Test Eval",
             eval_set_filter_id="tag::tag1",
             eval_configs_filter_id="tag::tag2",
@@ -633,3 +631,147 @@ def test_eval_run_eval_config_eval_validation():
             output="test output",
             scores={"score": 1.0},
         )
+
+
+@pytest.mark.parametrize(
+    "template_properties,should_raise,expected_error",
+    [
+        # Valid cases
+        (
+            {"issue_prompt": "Test issue prompt"},
+            False,
+            None,
+        ),
+        (
+            {
+                "issue_prompt": "Test issue prompt",
+                "failure_example": "Test failure example",
+            },
+            False,
+            None,
+        ),
+        (
+            {
+                "issue_prompt": "Test issue prompt",
+                "failure_example": "Test failure example",
+                "pass_example": "Test pass example",
+            },
+            False,
+            None,
+        ),
+        (
+            {
+                "issue_prompt": "",
+                "failure_example": "",
+                "pass_example": "",
+            },
+            False,
+            None,
+        ),
+        # Invalid cases
+        (
+            {},
+            True,
+            "issue_prompt is required for issue template",
+        ),
+        (
+            {"failure_example": "Test failure example"},
+            True,
+            "issue_prompt is required for issue template",
+        ),
+        (
+            {"issue_prompt": 123},
+            True,
+            "issue_prompt is required for issue template",
+        ),
+        (
+            {
+                "issue_prompt": "Test issue prompt",
+                "failure_example": 456,
+            },
+            True,
+            "failure_example is optional for issue template, but if provided must be a string",
+        ),
+        (
+            {
+                "issue_prompt": "Test issue prompt",
+                "failure_example": "Test failure example",
+                "pass_example": 789,
+            },
+            True,
+            "pass_example is optional for issue template, but if provided must be a string",
+        ),
+    ],
+)
+def test_eval_template_properties_issue_template_validation(
+    template_properties, should_raise, expected_error
+):
+    """Test issue template validation with various property combinations"""
+    if should_raise:
+        with pytest.raises(ValueError, match=expected_error):
+            Eval(
+                name="Test Eval",
+                template=EvalTemplateId.issue,
+                eval_set_filter_id="tag::tag1",
+                eval_configs_filter_id="tag::tag2",
+                output_scores=[
+                    EvalOutputScore(
+                        name="score",
+                        type=TaskOutputRatingType.pass_fail,
+                    )
+                ],
+                template_properties=template_properties,
+            )
+    else:
+        eval = Eval(
+            name="Test Eval",
+            template=EvalTemplateId.issue,
+            eval_set_filter_id="tag::tag1",
+            eval_configs_filter_id="tag::tag2",
+            output_scores=[
+                EvalOutputScore(
+                    name="score",
+                    type=TaskOutputRatingType.pass_fail,
+                )
+            ],
+            template_properties=template_properties,
+        )
+        assert eval.template == EvalTemplateId.issue
+        for key, value in template_properties.items():
+            assert eval.template_properties[key] == value
+
+
+@pytest.mark.parametrize(
+    "template,template_properties",
+    [
+        (EvalTemplateId.kiln_requirements, {"random_property": "random_value"}),
+        (EvalTemplateId.toxicity, {}),
+        (EvalTemplateId.bias, {"some_property": 123}),
+        (EvalTemplateId.maliciousness, {"test": True}),
+        (EvalTemplateId.factual_correctness, {"score": 4.5}),
+        (EvalTemplateId.jailbreak, {"prompt": "test"}),
+        (
+            None,
+            {"issue_prompt": "This should not be validated", "failure_example": 123},
+        ),
+    ],
+)
+def test_eval_template_properties_non_issue_templates(template, template_properties):
+    """Test that non-issue templates pass validation regardless of template_properties"""
+    eval = Eval(
+        name="Test Eval",
+        template=template,
+        eval_set_filter_id="tag::tag1",
+        eval_configs_filter_id="tag::tag2",
+        output_scores=[
+            EvalOutputScore(
+                name="score",
+                type=TaskOutputRatingType.pass_fail,
+            )
+        ],
+        template_properties=template_properties,
+    )
+
+    assert eval.template == template
+    for key, value in template_properties.items():
+        assert eval.template_properties[key] == value

kiln_ai/datamodel/test_model_perf.py

@@ -121,6 +121,6 @@ def test_benchmark_load_from_file(benchmark, task_run):
 
     # I get 8k ops per second on my MBP. Lower value here for CI and parallel testing.
     # Prior to optimization was 290 ops per second.
-    print(f"Ops per second: {ops_per_second:.6f}")
+    # sys.stdout.write(f"Ops per second: {ops_per_second:.6f}")
     if ops_per_second < 500:
         pytest.fail(f"Ops per second: {ops_per_second:.6f}, expected more than 1k ops")

kiln_ai/datamodel/test_prompt_id.py

@@ -29,7 +29,7 @@ def test_valid_saved_prompt_id():
 
 def test_valid_fine_tune_prompt_id():
     """Test that valid fine-tune prompt IDs are accepted"""
-    valid_id = "fine_tune_prompt::ft_123456"
+    valid_id = "fine_tune_prompt::project_123::task_456::ft_123456"
     model = ModelTester(prompt_id=valid_id)
     assert model.prompt_id == valid_id
 
@@ -53,6 +53,10 @@ def test_invalid_saved_prompt_id_format(invalid_id):
     [
         ("fine_tune_prompt::", "Invalid fine-tune prompt ID: fine_tune_prompt::"),
         ("fine_tune_prompt", "Invalid prompt ID: fine_tune_prompt"),
+        (
+            "fine_tune_prompt::ft_123456",
+            "Invalid fine-tune prompt ID: fine_tune_prompt::ft_123456",
+        ),
     ],
 )
 def test_invalid_fine_tune_prompt_id_format(invalid_id, expected_error):

kiln_ai/datamodel/test_task.py

@@ -323,3 +323,8 @@ def test_run_config_upgrade_old_entries():
     assert parsed.name == "test name"
     assert parsed.created_by == "scosman"
     assert parsed.run_config_properties.structured_output_mode == "unknown"
+
+
+def test_task_name_unicode_name():
+    task = Task(name="你好", instruction="Do something")
+    assert task.name == "你好"

kiln_ai/utils/config.py

@@ -124,6 +124,11 @@ class Config:
                 env_var="WANDB_API_KEY",
                 sensitive=True,
             ),
+            "siliconflow_cn_api_key": ConfigProperty(
+                str,
+                env_var="SILICONFLOW_CN_API_KEY",
+                sensitive=True,
+            ),
             "wandb_base_url": ConfigProperty(
                 str,
                 env_var="WANDB_BASE_URL",
@@ -137,6 +142,11 @@ class Config:
                 default_lambda=lambda: [],
                 sensitive_keys=["api_key"],
             ),
+            "cerebras_api_key": ConfigProperty(
+                str,
+                env_var="CEREBRAS_API_KEY",
+                sensitive=True,
+            ),
         }
         self._lock = threading.Lock()
         self._settings = self.load_settings()

kiln_ai/utils/logging.py

@@ -63,14 +63,14 @@ class CustomLiteLLMLogger(CustomLogger):
         # Print the formatted input data for the request in API format, pretty print
         try:
             self.logger.info(
-                f"Formatted Input Data (API):\n{json.dumps(data, indent=2)}"
+                f"Formatted Input Data (API):\n{json.dumps(data, indent=2, ensure_ascii=False)}"
             )
         except Exception as e:
             self.logger.info(f"Formatted Input Data (API): Could not print {e}")
 
         # Print the messages for the request in LiteLLM Message list, pretty print
         try:
-            json_messages = json.dumps(messages, indent=2)
+            json_messages = json.dumps(messages, indent=2, ensure_ascii=False)
             self.logger.info(f"Messages:\n{json_messages}")
         except Exception as e:
             self.logger.info(f"Messages: Could not print {e}")
@@ -115,7 +115,7 @@ class CustomLiteLLMLogger(CustomLogger):
                         # JSON format logs if possible
                         json_content = json.loads(content)
                         self.logger.info(
-                            f"Model Response Content:\n{json.dumps(json_content, indent=2)}"
+                            f"Model Response Content:\n{json.dumps(json_content, indent=2, ensure_ascii=False)}"
                         )
                     except Exception:
                         self.logger.info(f"Model Response Content:\n{content}")
@@ -149,6 +149,7 @@ def setup_litellm_logging(filename: str = "model_calls.log"):
         get_log_file_path(filename),
         maxBytes=5 * 1024 * 1024,  # 5MB
         backupCount=3,
+        encoding="utf-8",
     )
 
     # Set formatter to match the default formatting

{kiln_ai-0.17.0.dist-info → kiln_ai-0.19.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: kiln-ai
-Version: 0.17.0
+Version: 0.19.0
 Summary: Kiln AI
 Project-URL: Homepage, https://getkiln.ai
 Project-URL: Repository, https://github.com/Kiln-AI/kiln
@@ -19,7 +19,7 @@ Requires-Dist: boto3>=1.37.10
 Requires-Dist: coverage>=7.6.4
 Requires-Dist: google-cloud-aiplatform>=1.84.0
 Requires-Dist: jsonschema>=4.23.0
-Requires-Dist: litellm>=1.67.0
+Requires-Dist: litellm>=1.72.6
 Requires-Dist: openai>=1.53.0
 Requires-Dist: pdoc>=15.0.0
 Requires-Dist: pydantic>=2.9.2
@@ -65,6 +65,7 @@ The library has a [comprehensive set of docs](https://kiln-ai.github.io/Kiln/kil
 
 ## Table of Contents
 
+- [Connecting AI Providers](#connecting-ai-providers-openai-openrouter-ollama-etc)
 - [Using the Kiln Data Model](#using-the-kiln-data-model)
   - [Understanding the Kiln Data Model](#understanding-the-kiln-data-model)
   - [Datamodel Overview](#datamodel-overview)
@@ -73,6 +74,7 @@ The library has a [comprehensive set of docs](https://kiln-ai.github.io/Kiln/kil
   - [Using your Kiln Dataset in a Notebook or Project](#using-your-kiln-dataset-in-a-notebook-or-project)
   - [Using Kiln Dataset in Pandas](#using-kiln-dataset-in-pandas)
   - [Building and Running a Kiln Task from Code](#building-and-running-a-kiln-task-from-code)
+  - [Tagging Task Runs Programmatically](#tagging-task-runs-programmatically)
   - [Adding Custom Model or AI Provider from Code](#adding-custom-model-or-ai-provider-from-code)
 - [Full API Reference](#full-api-reference)
 
@@ -82,6 +84,12 @@ The library has a [comprehensive set of docs](https://kiln-ai.github.io/Kiln/kil
 pip install kiln-ai
 ```
 
+## Connecting AI Providers (OpenAI, OpenRouter, Ollama, etc)
+
+The easiest way to connect AI providers is to use the Kiln app UI. Once connected in the UI, credentials will be stored to `~/.kiln_ai/settings.yml`, which will be available to the library.
+
+For configuring credentials from code or connecting custom servers/model, see [Adding Custom Model or AI Provider from Code](#adding-custom-model-or-ai-provider-from-code).
+
 ## Using the Kiln Data Model
 
 ### Understanding the Kiln Data Model
@@ -179,7 +187,10 @@ item = kiln_ai.datamodel.TaskRun(
             type=kiln_ai.datamodel.DataSourceType.human,
             properties={"created_by": "Jane Doe"},
         ),
-        rating=kiln_ai.datamodel.TaskOutputRating(score=5,type="five_star"),
+        rating=kiln_ai.datamodel.TaskOutputRating(
+            value=5,
+            type=kiln_ai.datamodel.datamodel_enums.five_star,
+        ),
     ),
 )
 item.save_to_file()
@@ -270,6 +281,25 @@ for run in task.runs():
 
 ```
 
+## Tagging Task Runs Programmatically
+
+You can also tag your Kiln Task runs programmatically:
+
+```py
+# Load your Kiln Task from disk
+task_path = "/Users/youruser/Kiln Projects/test project/tasks/632780983478 - Joke Generator/task.kiln"
+task = kiln_ai.datamodel.Task.load_from_file(task_path)
+
+for run in task.runs():
+    # Parse the task output from JSON
+    output = json.loads(run.output.output)
+
+    # Add a tag if the punchline is unusually short
+    if len(output["punchline"]) < 100:
+        run.tags.append("very_short")
+        run.save_to_file()  # Persist the updated tags
+```
+
 ### Adding Custom Model or AI Provider from Code
 
 You can add additional AI models and providers to Kiln.

{kiln_ai-0.17.0.dist-info → kiln_ai-0.19.0.dist-info}/RECORD

@@ -1,113 +1,115 @@
 kiln_ai/__init__.py,sha256=Sc4z8LRVFMwJUoc_DPVUriSXTZ6PO9MaJ80PhRbKyB8,34
 kiln_ai/adapters/__init__.py,sha256=5GTN1bnEWz1NxWrlzsI6CfUio-D1jG74_-fx_PXJkY8,1027
-kiln_ai/adapters/adapter_registry.py,sha256=oEY6Zy6rtzs5mYLSZVhCzNWE4UJiUzIkVmmE_UYnBbE,8529
-kiln_ai/adapters/ml_model_list.py,sha256=kXt2eJ68HCUZ2dzfl9t-9pUv_s8149yMjRjmDi3GWuA,88641
+kiln_ai/adapters/adapter_registry.py,sha256=K5GdKwtfj4Cgpx1xzI1WS1uAxi12yhfUdjOq9gT7eWA,9725
+kiln_ai/adapters/ml_model_list.py,sha256=cURwa5ovOCkHxx5yCMMl5T7vclbUltaPyOBt__FIxg8,119552
 kiln_ai/adapters/ollama_tools.py,sha256=uObtLWfqKb9RXHN-TGGw2Y1FQlEMe0u8FgszI0zQn6U,3550
 kiln_ai/adapters/prompt_builders.py,sha256=R5IgZ7I2Ftx7i67xQb9UwcHE5gbPxgvgys_UAs3fc6A,15165
-kiln_ai/adapters/provider_tools.py,sha256=HIGsU38cAWKGONi8-Rh66WNOved3O0w9OYbj9tU_HU4,16794
+kiln_ai/adapters/provider_tools.py,sha256=gLRG3s2UTZ_DCOQILXuhTAMmcVgj6e5sIIPPIKZ7dKk,17217
+kiln_ai/adapters/remote_config.py,sha256=qXRjXVoF_f31XomSGdOdFyZL-_fNO66rp8JK5Wji1PE,3954
 kiln_ai/adapters/run_output.py,sha256=RAi2Qp6dmqJVNm3CxbNTdAuhitHfH5NiUGbf6ygUP-k,257
-kiln_ai/adapters/test_adapter_registry.py,sha256=W4KoZ5cFTZwmW3i19cCgwzmcG802tpFC0GrUnZcD8_4,8378
-kiln_ai/adapters/test_ml_model_list.py,sha256=L5XdRz6c2vbfSYe0rWfsNOM73lZbuGsnG1gzgVPEUA8,6726
+kiln_ai/adapters/test_adapter_registry.py,sha256=QofH8Gzx7tWk3wvzf7HL476xh9GGqwabFmoWZUDE4rA,9415
+kiln_ai/adapters/test_ml_model_list.py,sha256=yoJUTd7TxDU2G_mJmZqX46e0EZgE7z60c71Anm0RiCw,7940
 kiln_ai/adapters/test_ollama_tools.py,sha256=xAUzL0IVmmXadVehJu1WjqbhpKEYGAgGt3pWx7hrubc,2514
-kiln_ai/adapters/test_prompt_adaptors.py,sha256=nGq6H84rDMl72WUP2xqorQYW5yot0Z6sTQPXRL0mh5Y,8291
+kiln_ai/adapters/test_prompt_adaptors.py,sha256=ikEeaWvBI4OQ_bVRYkUGPtjZIz37Ej2P8bAr85DGYNQ,8180
 kiln_ai/adapters/test_prompt_builders.py,sha256=SqGkluKZd7m7SPuq64NAprG0db11RDqoYai1m-1QgeQ,22585
-kiln_ai/adapters/test_provider_tools.py,sha256=QCknT944HN7jKLeSi9qS0cH_O5jKMMLThX7kWfAcJqg,32090
+kiln_ai/adapters/test_provider_tools.py,sha256=Qts7nt8_9_WP0LPVlEJcwswcbqtVNtIa6u0IGpeQO1Y,32448
+kiln_ai/adapters/test_remote_config.py,sha256=RxAJtTtzonum9LW0nT5HMcwCvhgSZLqjFs8a9gbJkEY,17095
 kiln_ai/adapters/chat/__init__.py,sha256=ELydKUGeCcTiBJogzZUT8VXqr5kDtmoB8_GoyO28qR0,192
-kiln_ai/adapters/chat/chat_formatter.py,sha256=aBGpMUUBx0V_-g5GXkAhmhIqIyfMwd7kphFu6E8ivTM,8343
+kiln_ai/adapters/chat/chat_formatter.py,sha256=lr2zIj-jbZNStaM1VZSvoaN_e6KDKNWiNNPC6RDQ6ug,8321
 kiln_ai/adapters/chat/test_chat_formatter.py,sha256=MVEZTSIFBwLvplOmit-4TDdcmPXsaMZMQEwoXWmq1FI,4603
 kiln_ai/adapters/data_gen/__init__.py,sha256=QTZWaf7kq5BorhPvexJfwDEKmjRmIbhwW9ei8LW2SIs,276
-kiln_ai/adapters/data_gen/data_gen_prompts.py,sha256=kudjHnAz7L3q0k_NLyTlaIV7M0uRFrxXNcfcnjOE2uc,5810
-kiln_ai/adapters/data_gen/data_gen_task.py,sha256=0PuYCcj09BtpgNj23mKj_L45mKZBdV5VreUeZ-Tj_xM,6642
-kiln_ai/adapters/data_gen/test_data_gen_task.py,sha256=_v80CcPpTRQRBGgO9U-tj8PNg5Ixzc-7xCgSocCRai0,10706
+kiln_ai/adapters/data_gen/data_gen_prompts.py,sha256=06gqU7uJDBr-G-txeGV9a4_NgKlbxDxhcylWclcVVxg,8792
+kiln_ai/adapters/data_gen/data_gen_task.py,sha256=El-5Dad65JvYkgr9Q40LfdAQ7IUx9BIjxF5c_j8sEq0,7121
+kiln_ai/adapters/data_gen/test_data_gen_task.py,sha256=nIuyJkYWsg9VUaLh95csrx8CnA3d8ivwmPjqTnOoz4Q,21308
 kiln_ai/adapters/eval/__init__.py,sha256=0ptbK0ZxWuraxGn_WMgmE1tcaq0k5t-g-52kVohvWCg,693
-kiln_ai/adapters/eval/base_eval.py,sha256=Ms-OPfOmp4_7jIgAxi9QbwEPQYHkkHA-F2ZW11S6MLg,7324
-kiln_ai/adapters/eval/eval_runner.py,sha256=EjgAM88RhV51aIM1tNwmtnOU8a7uyaSM95ZYu0JZ-Pc,8791
-kiln_ai/adapters/eval/g_eval.py,sha256=yDsyepDhvsi67F2gi85KZQO_0UoWKNJSpxc43ufvJOU,16098
+kiln_ai/adapters/eval/base_eval.py,sha256=g7BWaljeuRquP-ygOBO6jwwzuWyNTYy_QhF0lxZlr-A,7476
+kiln_ai/adapters/eval/eval_runner.py,sha256=W3qBVWBhKQzgxW2jp4ht_WALyjqwDJnsbX9L_85XUDc,8996
+kiln_ai/adapters/eval/g_eval.py,sha256=NVq6iXgKbOG8qimivTciEh2C6lmgrKFXcbNu9yBBeSQ,16304
 kiln_ai/adapters/eval/registry.py,sha256=gZ_s0VgEx79Fswkgi1tS4yOl7lzpkvUBJZ62RldhM_w,626
-kiln_ai/adapters/eval/test_base_eval.py,sha256=zNoIrFqImgccOazC4yNof11ZObHkvh_DVnaaBBNw_jU,15776
-kiln_ai/adapters/eval/test_eval_runner.py,sha256=37cmZPjvZaNlNDzmstmOEUwQEFhpafD9jRhDD1UH6PE,18718
-kiln_ai/adapters/eval/test_g_eval.py,sha256=KaVjG6aHQQzGlOd4J_7S8YKR3EmA2aFpY8VOEDdrRHY,16479
-kiln_ai/adapters/eval/test_g_eval_data.py,sha256=8caiZfLWnXVX8alrBPrH7L7gqqSS9vO7u6PzcHurQcA,27769
+kiln_ai/adapters/eval/test_base_eval.py,sha256=JRts5jsvD2XMqdegsenVl2yy_sL6ru9W7n9wPEbKD7U,15994
+kiln_ai/adapters/eval/test_eval_runner.py,sha256=vSDlYo_ujB1EReCLUZlcnqNaamG2wB9QRil_X_oTEm0,18587
+kiln_ai/adapters/eval/test_g_eval.py,sha256=4sn0sfZdKHgSW_Y-fbZAE88JAUERl93MhlZxKFkCMiY,21002
+kiln_ai/adapters/eval/test_g_eval_data.py,sha256=9IIQE6YNmf-epogQapJloocTMSpF4dkTIeEEsgF474U,27805
 kiln_ai/adapters/fine_tune/__init__.py,sha256=DxdTR60chwgck1aEoVYWyfWi6Ed2ZkdJj0lar-SEAj4,257
-kiln_ai/adapters/fine_tune/base_finetune.py,sha256=CGho356kKnxQsvkHws9XfWuBYV0nmbCvEkff21D4e48,5914
-kiln_ai/adapters/fine_tune/dataset_formatter.py,sha256=yOVefk2qXwX29Y2nKU0XWtQgf1kNNMY5M0NhzsaatZ8,12975
+kiln_ai/adapters/fine_tune/base_finetune.py,sha256=4mLrFHvERU76_e8bCuOGLTc5UAdFn4UUCuPAe_3uD0Q,5875
+kiln_ai/adapters/fine_tune/dataset_formatter.py,sha256=oWYXOdkRaVZR_mIeox2gLf5_LZJ79hHV9PThk82zHtU,12928
 kiln_ai/adapters/fine_tune/finetune_registry.py,sha256=9RJLjviSoN3dQnKJE9Ss7df7dtdJgbuShB8IUcI-q9k,726
-kiln_ai/adapters/fine_tune/fireworks_finetune.py,sha256=ze0QxghpHAqwO9nXOTkDEC9irmqduX5bjIhZDU0DCZQ,20101
+kiln_ai/adapters/fine_tune/fireworks_finetune.py,sha256=WuC2zkZQTD6msew81Dd6GMT6OV5mAl33_K9K7Os_OBY,20833
 kiln_ai/adapters/fine_tune/openai_finetune.py,sha256=Dz9E_0BWfrIkvv8ArZe-RKPwbIKPZ3v8rfbc3JELyTY,8571
 kiln_ai/adapters/fine_tune/test_base_finetune.py,sha256=atZI54SC6mVili5dEIuqwv2groLCZWYKSMzDZYjbUxQ,10705
-kiln_ai/adapters/fine_tune/test_dataset_formatter.py,sha256=7-axt5WPk1cEmsojr9xG6o1PHu7EvEn_YWIg1ZmqRK0,29907
-kiln_ai/adapters/fine_tune/test_fireworks_tinetune.py,sha256=dzOOFIzgRUGGNlfA7-L1URI4qwMVHg1fuUc8RBqrDNI,36765
+kiln_ai/adapters/fine_tune/test_dataset_formatter.py,sha256=OmF3jW78FBHvDes6lusY8hWgLLHz5XBgHdDVWv2ccAI,29898
+kiln_ai/adapters/fine_tune/test_fireworks_tinetune.py,sha256=Xq89amS0VyYP8zi8YteTudYHhykEK7-BZXOsIo_7uKE,37146
 kiln_ai/adapters/fine_tune/test_openai_finetune.py,sha256=NtMFo3LZjpAsNHFvIMHsFmWRrHrOKjniRvrZTx4nXhg,20001
 kiln_ai/adapters/fine_tune/test_together_finetune.py,sha256=ZwYmT4pPkhU-45Q-5fs4WL16dQ4OyvI7KXPevsrA24E,18122
-kiln_ai/adapters/fine_tune/test_vertex_finetune.py,sha256=W7oz6ZJktde1WludBb2lFpTXmU7VQN1FkeJpyN1dPbI,19218
-kiln_ai/adapters/fine_tune/together_finetune.py,sha256=rsJETfwkx21UzJH1CUd_hP-v1zQNBxK5b2uJIYSTpS4,14775
+kiln_ai/adapters/fine_tune/test_vertex_finetune.py,sha256=DBQbNUgpqDEvjjk7eHwBvDXPCEkhf4yfUw8QPi5fFPE,19178
+kiln_ai/adapters/fine_tune/together_finetune.py,sha256=KpJBZt0NBNUNLsVEYYC4Dzob1jnlOFQnVI8Q-4kY9d8,14766
 kiln_ai/adapters/fine_tune/vertex_finetune.py,sha256=Ik6Ov711-oruJnMHpVZTPimWJY2W_JnfdKIdR2djGrc,8545
 kiln_ai/adapters/model_adapters/__init__.py,sha256=m5GRtOHwVVvp_XDOss8c1X3NFf1wQQlC2eBgI4tXQhM,212
-kiln_ai/adapters/model_adapters/base_adapter.py,sha256=mGF3ozrFp14_elYXghgzZiWrEKo6EGY9lSdo3aRaJ-A,13101
-kiln_ai/adapters/model_adapters/litellm_adapter.py,sha256=G9SBxw-B6uB8I8LyKaZ6WDAq2OaUOCmvoVNYKbNdVu4,17074
+kiln_ai/adapters/model_adapters/base_adapter.py,sha256=9z6i72nX5xIP9cvHFhz47sXl_g2Gv9DPnFjemQUiRSw,13287
+kiln_ai/adapters/model_adapters/litellm_adapter.py,sha256=TGCgCKxJhMEFSgmVRDXwW5OlRnfP8omAAfv2dHMhgks,17938
 kiln_ai/adapters/model_adapters/litellm_config.py,sha256=zOQEkYKeoQ0FIbxTkyyoaGtaQiB9eYK3IuyUgqSwzLE,485
-kiln_ai/adapters/model_adapters/test_base_adapter.py,sha256=CWvVCr3Y_m8iQ1jU60PDG7qPpIP3gOSAXUfOUsQVAfI,16050
-kiln_ai/adapters/model_adapters/test_litellm_adapter.py,sha256=NZ9rXUHCDjNNuvzMjOHmmT5U02HXCm7WkuyNOXYaLNU,17301
+kiln_ai/adapters/model_adapters/test_base_adapter.py,sha256=POHvlGTdZGCZf_ktrds3oBbvT06t5Gp-iGBEGFFN3sA,15958
+kiln_ai/adapters/model_adapters/test_litellm_adapter.py,sha256=28xtjin8RZnf49-3RATtvon05mE4OxTxv0yd0ct1Dhc,19663
 kiln_ai/adapters/model_adapters/test_saving_adapter_results.py,sha256=xkd_b9GS-6ybst97FsS_XrvyTXMlEufYF1Q8CGpE1V8,9697
-kiln_ai/adapters/model_adapters/test_structured_output.py,sha256=-4NWrDTYJXC53UKoKfYzAg-uyD73I-iApqN-9FNB2qM,13463
+kiln_ai/adapters/model_adapters/test_structured_output.py,sha256=1yYhQDchn3rmC44FMng55ViC-uBmNrqXUxuhwX-CWkc,13860
 kiln_ai/adapters/parsers/__init__.py,sha256=TGJS_8JhjUwg5Bnq4cDmwt5eIRo4vowmcL2A72L1Hzk,202
 kiln_ai/adapters/parsers/base_parser.py,sha256=AE8UYCttmVXbilpICotnDdFYTFhGMiBJHrxIUgPTAWM,280
 kiln_ai/adapters/parsers/json_parser.py,sha256=IszrBrhIFrrVr76UZsuejkBdqpZG27mU72264HVgVzE,1274
-kiln_ai/adapters/parsers/parser_registry.py,sha256=Lna-uYHBigKeq9e-mUbNQKcsZM1M7cDvxu7Gv4ECjLI,735
-kiln_ai/adapters/parsers/r1_parser.py,sha256=ucDRHPoFrmr1eOSNFEG0won6TPxkCKfGJQNNJng5qHA,3626
+kiln_ai/adapters/parsers/parser_registry.py,sha256=BbY12zbXj-FgdJ8WqEcBLg26WXoDk2fmriLRo2LYCSk,710
+kiln_ai/adapters/parsers/r1_parser.py,sha256=fbO-Ab2CVj6K756GaLGwNL6kkNqBBq7dWHtenL1oiNI,3559
 kiln_ai/adapters/parsers/request_formatters.py,sha256=NdZO8zcZy9tkea8JaD5c_OeoeVjoYYTG0GjwF812STw,1124
 kiln_ai/adapters/parsers/test_json_parser.py,sha256=9kdWe_vRC5wjP8A1Ym6Zu6enDIz4ARCNiRpcZr7_3ak,1971
 kiln_ai/adapters/parsers/test_parser_registry.py,sha256=FFJQgaKVu67yK4W7w_b26tuEYPPYGppfttJ0y5ctWUo,1041
-kiln_ai/adapters/parsers/test_r1_parser.py,sha256=XbB0saThkcgOvoYNq3TxDgILUu5Me38yeKoQonqaN-g,6296
+kiln_ai/adapters/parsers/test_r1_parser.py,sha256=gj4N_JZmyBwZRqPDiGY0j30qQ1mrQhGrTuCRAgQv6_M,6306
 kiln_ai/adapters/parsers/test_request_formatters.py,sha256=t3FlKT_Tte2u8zXJTMl8VaE8IrSzumuBysahbGesrbU,2090
 kiln_ai/adapters/repair/__init__.py,sha256=dOO9MEpEhjiwzDVFg3MNfA2bKMPlax9iekDatpTkX8E,217
-kiln_ai/adapters/repair/repair_task.py,sha256=iW0bHWQq6Tir6ULTATWFS0zpwNji8Tbwhm2lZu52RsM,3342
+kiln_ai/adapters/repair/repair_task.py,sha256=ziOKNo2dT0a4ggLIDk5ETfSTbuxn_xgI60w1Ajd0sww,3286
 kiln_ai/adapters/repair/test_repair_task.py,sha256=fMlad29edA2tTt4t7cv6cXVWhuWOJ9x1Mpb3wJFTD1I,8603
 kiln_ai/datamodel/__init__.py,sha256=eHDUB9ltKmnsx0TAX310o5OMLnZbSznMJLFRnNk6HlA,1927
-kiln_ai/datamodel/basemodel.py,sha256=Qsr3Kxq5B5rfPZxQFrfOMTGwInnfYy5srd8nK0nwOs0,22204
-kiln_ai/datamodel/datamodel_enums.py,sha256=7BpyiXEtNsz43el7zTAeaaozDVUZUJCFKHsoaXoTomQ,3796
+kiln_ai/datamodel/basemodel.py,sha256=jyB351QSq6ll9JwzxCBaCY1aXNRp5nDF1dBdhVA6KRM,23871
+kiln_ai/datamodel/datamodel_enums.py,sha256=jTICOyCIUJ3hr13T2_CXPLzxgLtSx3JlE4GJ2vPnVKI,3860
 kiln_ai/datamodel/dataset_filters.py,sha256=zXS5QeCqWa1vAxs3Pyy5FJtySx9yYJzxsCMANd23-3c,5502
-kiln_ai/datamodel/dataset_split.py,sha256=GDF3Pj3NLG42X8kjR606R_aN14rpqx8Ig56FG8NCn9k,5790
-kiln_ai/datamodel/eval.py,sha256=larT3emjpLVYOSB2ga5GHTS4O7l9mJwtIO4EqCyYF7g,14214
-kiln_ai/datamodel/finetune.py,sha256=mOTM0KEbRNZ5fvzgpUtLRugaGkhWQKFwdsQWggIzMls,4813
+kiln_ai/datamodel/dataset_split.py,sha256=dAqwwNtWhKntEGJtauT0-YDG8aLFISelYS7o8TWG0TE,5923
+kiln_ai/datamodel/eval.py,sha256=xYRSUGxINR-4pZvAUkg8GpemWFlyywtkA707oT3R6OM,15897
+kiln_ai/datamodel/finetune.py,sha256=3cwg9FTbhYDH2DEWS_L8lMS5KPok9mJItWjYHC6LnH0,4835
 kiln_ai/datamodel/json_schema.py,sha256=o50wSp8frRXjT-NZjml4-Is7LNoF7DQP4g3AaaYzBfI,3379
 kiln_ai/datamodel/model_cache.py,sha256=9X4aAigbkFdytckgw8InCMh86uBna0ME_1HJSeMPEn0,4495
-kiln_ai/datamodel/project.py,sha256=uVH2_3TDFtsG_tpts81A-zbd9uPDFxAwMCKZt_km3IE,727
-kiln_ai/datamodel/prompt.py,sha256=70JPYHfgyX18cHW_DXoMzIOA28Jbaz6gyabElmpycyc,1161
-kiln_ai/datamodel/prompt_id.py,sha256=S4Wcrt05USN-JrO8BeDlNYGrcoTOVocR3iUxBbgoq1c,2566
+kiln_ai/datamodel/project.py,sha256=xdO0S_pnF1uoTr-DfuF1Xr1tK8WrhbHvXKk6AhfrjB0,778
+kiln_ai/datamodel/prompt.py,sha256=ZvqbycqP3XuEF7lXqCPTXlDHrS0OPZOUhU5T95_ns6M,1210
+kiln_ai/datamodel/prompt_id.py,sha256=oBnU_j8g3Xnj8TJbgQAQfW7I4vcyna-pDxFbiLvPs74,2631
 kiln_ai/datamodel/registry.py,sha256=XwGFXJFKZtOpR1Z9ven6SftggfADdZRm8TFxCEVtfUQ,957
 kiln_ai/datamodel/strict_mode.py,sha256=sm4Xka8mnJHCShtbh6MMU5dDQv-cLj8lHgHkmFKpsl0,849
-kiln_ai/datamodel/task.py,sha256=ZjTfOxk4Op83CAVp1KFsuHeaeqBNAvdVS-QQCKCNIug,9253
-kiln_ai/datamodel/task_output.py,sha256=P7bRI-Qn2n7f8XpR7hwgSTKhx3gzvsZ1aJOQMl-YW4s,13245
-kiln_ai/datamodel/task_run.py,sha256=0Lwc7zQoeWzuvHYv00lUGfsNQp8gCKmCVKAq3FDrruY,8806
-kiln_ai/datamodel/test_basemodel.py,sha256=1__3dVyxCVMQH7jhBYYEYgaCgtc08faSIlVOLfPdMZ0,18021
+kiln_ai/datamodel/task.py,sha256=rKwPN7xzph94vYvqepTuBpl9mi8E_665i6EwbLtnKDU,9412
+kiln_ai/datamodel/task_output.py,sha256=404qa5jQhdDUzcYj4043v8YERhoHc61AVdaDGsDCY1w,13193
+kiln_ai/datamodel/task_run.py,sha256=rTPqmN0VOiw7P8Fy0cVQXzV7h0sO2-SWqhGsYN0owLU,8759
+kiln_ai/datamodel/test_basemodel.py,sha256=ssa8jZKJHeew2Xx0k9fUPKaCcw-VQKioACNbbPVJo1o,25254
 kiln_ai/datamodel/test_dataset_filters.py,sha256=TFHQZLB0rJxnxsFjD546eXVFxZNAJi0sPZ8j24wYd1w,5322
 kiln_ai/datamodel/test_dataset_split.py,sha256=mXB8udml_6U4BXR0xAGHsFINnhdcTDB1qhuZbQemk-w,11055
 kiln_ai/datamodel/test_datasource.py,sha256=H4Kc-Im9eM7WnADWZXdoiOIrOl05RtkyuhTCKiRimyU,3905
-kiln_ai/datamodel/test_eval_model.py,sha256=J7MqwWBgPpeXGqh3IacVUUHdZFJSZ2MgTsUNu-hNOJw,19528
+kiln_ai/datamodel/test_eval_model.py,sha256=WbGWiIP4b6pu-Z117iVNsPLnrpFW2-sj89oOIQ8RLEk,23940
 kiln_ai/datamodel/test_example_models.py,sha256=dwLAAOXLvdKupE5Q1m6VNcHtfdhpi2qWtoEbC0nfJg8,26156
 kiln_ai/datamodel/test_json_schema.py,sha256=R0Cfc9WbieMslgvYsj2HFx8RHIq2fF9NcT5jH-kEqh4,4793
 kiln_ai/datamodel/test_model_cache.py,sha256=Fy-ucYNzS5JEG-8SFY4nVHA8iRbXXxai20f8_oGl97o,8184
-kiln_ai/datamodel/test_model_perf.py,sha256=9_76f__5XtZSHSjuaoiHRj2t-z3OWn-sSA4S9kH1jpY,3306
+kiln_ai/datamodel/test_model_perf.py,sha256=JcJpedvgK49zj_VIUc5DW0QjuIXhaBQbBpccUJVy2hU,3319
 kiln_ai/datamodel/test_models.py,sha256=wENuBiZ2Y-N_puFkitNZ1T-lBpVgQWotfhGug6k4AMY,21674
 kiln_ai/datamodel/test_nested_save.py,sha256=xciCddqvPyKyoyjC5Lx_3Kh1t4LJv1xYRAPazR3SRcs,5588
 kiln_ai/datamodel/test_output_rating.py,sha256=zvPIp2shAgCs2RQBgwYoL09fRA3krHvgAqUa91RlWR0,15125
-kiln_ai/datamodel/test_prompt_id.py,sha256=ihyXVPQi0dSLGnBM7rTXRnVaiWXhh7HJmSy4nZZKmso,4225
+kiln_ai/datamodel/test_prompt_id.py,sha256=_vcuiDvai8TENb6agk0F9ZsaTrAgoD1VkMxRIT2p9mQ,4384
 kiln_ai/datamodel/test_registry.py,sha256=PhS4anLi5Bf_023obuTlO5DALhtPB8WIc_bX12Yg6Po,2705
-kiln_ai/datamodel/test_task.py,sha256=X85YgGt7Y9kuv6-jE9kl5b8mPz3cgrEFqiZAZFWfpO0,11890
+kiln_ai/datamodel/test_task.py,sha256=LdMb2C8emK5vvgE1ACRFRmetWHgXGFgk_dstbgRhO54,12019
 kiln_ai/utils/__init__.py,sha256=PTD0MwBCKAMIOGsTAwsFaJOusTJJoRFTfOGqRvCaU-E,142
 kiln_ai/utils/async_job_runner.py,sha256=1gjoEq5yc2MOVjDo05O1wztguEuMC6l3haDZsltlvuw,3457
-kiln_ai/utils/config.py,sha256=Pk9w7C985jinE9_SDJzgkG1CkPtIaitLZqia_dtjhwQ,8520
+kiln_ai/utils/config.py,sha256=GRqq-52F4_ICuWiJzzaIJ-Z77ayqgLbIw3-Gq86zMUY,8852
 kiln_ai/utils/dataset_import.py,sha256=EqBBBopCEUy1JH4-EAsBETwGp4MFjzZGfUUBZ6FLfGY,9011
 kiln_ai/utils/exhaustive_error.py,sha256=TkkRixIAR3CPEKHeAJzyv0mtxp6BxUBKMvobA3vzQug,262
 kiln_ai/utils/formatting.py,sha256=VtB9oag0lOGv17dwT7OPX_3HzBfaU9GsLH-iLete0yM,97
-kiln_ai/utils/logging.py,sha256=LY7EnxZIsSuVxKP53JMLp-TdFomOMkxN793trN_kAws,6729
+kiln_ai/utils/logging.py,sha256=ixtv2Mm6-XWB2PVUPAWVRXnPJULvoTDrsp5NTAk3yp0,6815
 kiln_ai/utils/name_generator.py,sha256=v26TgpCwQbhQFcZvzgjZvURinjrOyyFhxpsI6NQrHKc,1914
 kiln_ai/utils/test_async_job_runner.py,sha256=8AwPfOlR09qzfhVm-djpkmuoyHxJJ19QEzs0WV5KFSQ,6813
 kiln_ai/utils/test_config.py,sha256=8goGdVLOO1OiSPnuDLUHbEJWJ8hAjaAIDg0myhR2A00,9687
 kiln_ai/utils/test_dataset_import.py,sha256=BEl38D95HQYpc7_jeB1N-ocOnKM1DLutp669cNrVOuE,25765
 kiln_ai/utils/test_name_geneator.py,sha256=9-hSTBshyakqlPbFnNcggwLrL7lcPTitauBYHg9jFWI,1513
-kiln_ai-0.17.0.dist-info/METADATA,sha256=W9UIYyQGlhO3fcDr9IBaxUTST_lF_-DHKGm3cCUPcH0,12263
-kiln_ai-0.17.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-kiln_ai-0.17.0.dist-info/licenses/LICENSE.txt,sha256=_NA5pnTYgRRr4qH6lE3X-TuZJ8iRcMUi5ASoGr-lEx8,1209
-kiln_ai-0.17.0.dist-info/RECORD,,
+kiln_ai-0.19.0.dist-info/METADATA,sha256=0PTv8ksKamQEk30IKehRIuIQ7TK5LfI0fANyjg8WjKc,13507
+kiln_ai-0.19.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+kiln_ai-0.19.0.dist-info/licenses/LICENSE.txt,sha256=_NA5pnTYgRRr4qH6lE3X-TuZJ8iRcMUi5ASoGr-lEx8,1209
+kiln_ai-0.19.0.dist-info/RECORD,,