python-flexeval 0.2.0__py3-none-any.whl → 0.3.0__py3-none-any.whl

flexeval/__about__.py

@@ -1 +1 @@
-__version__ = "0.2.0"
+__version__ = "0.3.0"

flexeval/classes/jsonview.py

@@ -0,0 +1,107 @@
+import json
+from collections import UserDict
+
+
+class JsonViewDict(UserDict):
+    """Dictionary that syncs changes back to the model field."""
+
+    def __init__(
+        self,
+        model_instance,
+        text_field_attr_name,
+        json_dumps_fn=json.dumps,
+        json_loads_fn=json.loads,
+    ):
+        self.model_instance = model_instance
+        self.text_field_attr_name = text_field_attr_name
+        self.json_dumps_fn = json_dumps_fn
+        self.json_loads_fn = json_loads_fn
+
+        text_value = getattr(model_instance, text_field_attr_name)
+        initial_data = self.json_loads_fn(text_value)
+        super().__init__(initial_data)
+
+    def _sync_to_model(self):
+        """Sync the current data back to the model field."""
+        json_str = self.json_loads_fn(self.data)
+        setattr(self.model_instance, self.text_field_attr_name, json_str)
+
+    # Override mutating methods to trigger sync
+    def __setitem__(self, key, value):
+        super().__setitem__(key, value)
+        self._sync_to_model()
+
+    def __delitem__(self, key):
+        super().__delitem__(key)
+        self._sync_to_model()
+
+    def clear(self):
+        super().clear()
+        self._sync_to_model()
+
+    def pop(self, key, *args):
+        result = super().pop(key, *args)
+        self._sync_to_model()
+        return result
+
+    def popitem(self):
+        result = super().popitem()
+        self._sync_to_model()
+        return result
+
+    def setdefault(self, key, default=None):
+        result = super().setdefault(key, default)
+        self._sync_to_model()
+        return result
+
+    def update(self, *args, **kwargs):
+        super().update(*args, **kwargs)
+        self._sync_to_model()
+
+
+class JsonView:
+    """Descriptor that provides dict-like access to a JSON text field.
+
+    Example:
+    class SomeModel(pw.Model):
+        some_field = pw.TextField(default="{}")
+        some_field_dict = JsonView(text_field_attr_name="some_field")
+
+    m = SomeModel()
+    m.some_field_dict["chosen_mistake"] = "whatever"
+    """
+
+    def __init__(self, text_field_attr_name):
+        self.text_field_attr_name = text_field_attr_name
+        self.attr_name = None
+
+    def __set_name__(self, owner, name):
+        """Called when the descriptor is assigned to a class attribute."""
+        self.attr_name = f"_{name}_dict"
+
+    def __get__(self, instance, owner):
+        if instance is None:
+            return self
+
+        # Check if we already have a cached JsonViewDict
+        if not hasattr(instance, self.attr_name):
+            if not hasattr(instance, self.text_field_attr_name):
+                raise ValueError(
+                    f"Failed to link this JsonView to field '{self.text_field_attr_name}' because it doesn't exist on this model instance."
+                )
+            # Cache a new JsonViewDict
+            json_dict = JsonViewDict(instance, self.text_field_attr_name)
+            setattr(instance, self.attr_name, json_dict)
+
+        return getattr(instance, self.attr_name)
+
+    def __set__(self, instance, value):
+        """Allow setting the entire dict."""
+        if isinstance(value, dict):
+            json_dict = JsonViewDict(instance, self.text_field_attr_name)
+            json_dict.update(value)
+            setattr(instance, self.attr_name, json_dict)
+        else:
+            raise ValueError(
+                f"This JsonView must be a dictionary to set linked field '{self.text_field_attr_name}' correctly."
+            )

flexeval/classes/message.py

@@ -10,6 +10,7 @@ from flexeval.classes.dataset import Dataset
 from flexeval.classes.eval_set_run import EvalSetRun
 from flexeval.classes.thread import Thread
 from flexeval.classes.turn import Turn
+from flexeval.classes.jsonview import JsonView
 from flexeval.configuration import completion_functions
 
 logger = logging.getLogger(__name__)
@@ -34,6 +35,10 @@ class Message(BaseModel):
     content = pw.TextField()
     context = pw.TextField(null=True)  # Previous messages
 
+    # metadata
+    metadata = pw.TextField(default="{}", null=False)
+    metadata_dict = JsonView("metadata")
+
     # helpers
     system_prompt = pw.TextField(null=True)
     is_flexeval_completion = pw.BooleanField(null=True)

flexeval/classes/thread.py

@@ -3,6 +3,7 @@ import peewee as pw
 from flexeval.classes.base import BaseModel
 from flexeval.classes.dataset import Dataset
 from flexeval.classes.eval_set_run import EvalSetRun
+from flexeval.classes.jsonview import JsonView
 
 
 class Thread(BaseModel):
@@ -20,6 +21,9 @@ class Thread(BaseModel):
 
     system_prompt = pw.TextField(null=True)
 
+    metadata = pw.TextField(default="{}", null=False)
+    metadata_dict = JsonView("metadata")
+
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
         self.metrics_to_evaluate = []

flexeval/data_loader.py

@@ -54,18 +54,13 @@ def load_jsonl(
                 max(1, nb_evaluations_per_thread)
             ):  # duplicate stored threads for averaged evaluation results
                 if thread_id in selected_thread_ids:
-                    thread_object = Thread.create(
-                        evalsetrun=dataset.evalsetrun,
-                        dataset=dataset,
-                        jsonl_thread_id=thread_id,
-                        eval_run_thread_id=str(thread_id)
-                        + "_"
-                        + str(thread_eval_run_id),
-                    )
+                    thread_json = json.loads(thread)
+                    # extract any metadata
+                    thread_metadata = thread_json.copy()
+                    del thread_metadata["input"]
 
-                    # Context
                     context = []
-                    thread_input = json.loads(thread)["input"]
+                    thread_input = thread_json["input"]
 
                     # Get system prompt used in the thread - assuming only 1
                     for message in thread_input:
@@ -78,15 +73,35 @@ def load_jsonl(
                         # Add the system prompt as context
                         context.append({"role": "system", "content": system_prompt})
 
+                    thread_object: Thread = Thread.create(
+                        evalsetrun=dataset.evalsetrun,
+                        dataset=dataset,
+                        jsonl_thread_id=thread_id,
+                        eval_run_thread_id=str(thread_id)
+                        + "_"
+                        + str(thread_eval_run_id),
+                        system_prompt=system_prompt,
+                        metadata=json.dumps(thread_metadata),
+                    )
+
                     # Create messages
                     index_in_thread = 0
                     for message in thread_input:
+                        if not isinstance(message, dict):
+                            raise ValueError(
+                                f"Can't load unknown object type; expected dict. Check JSONL format: {message}"
+                            )
                         role = message.get("role", None)
                         if role != "system":
                             # System message shouldn't be added as a separate message
                             system_prompt_for_this_message = ""
                             if role != "user":
                                 system_prompt_for_this_message = system_prompt
+                            message_metadata = message.copy()
+                            if "content" in message_metadata:
+                                del message_metadata["content"]
+                            if "role" in message_metadata:
+                                del message_metadata["role"]
                             Message.create(
                                 evalsetrun=dataset.evalsetrun,
                                 dataset=dataset,
@@ -95,9 +110,9 @@ def load_jsonl(
                                 role=role,
                                 content=message.get("content", None),
                                 context=json.dumps(context),
-                                metadata=message.get("metadata", None),
                                 is_flexeval_completion=False,
                                 system_prompt=system_prompt_for_this_message,
+                                metadata=json.dumps(message_metadata),
                             )
                             # Update context
                             context.append(

{python_flexeval-0.2.0.dist-info → python_flexeval-0.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-flexeval
-Version: 0.2.0
+Version: 0.3.0
 Summary: FlexEval is a tool for designing custom metrics, completion functions, and LLM-graded rubrics for evaluating the behavior of LLM-powered systems.
 Project-URL: Homepage, https://digitalharborfoundation.github.io/FlexEval/
 Project-URL: GitHub, https://github.com/DigitalHarborFoundation/FlexEval

{python_flexeval-0.2.0.dist-info → python_flexeval-0.3.0.dist-info}/RECORD

@@ -1,11 +1,11 @@
-flexeval/__about__.py,sha256=Zn1KFblwuFHiDRdRAiRnDBRkbPttWh44jKa5zG2ov0E,22
+flexeval/__about__.py,sha256=VrXpHDu3erkzwl_WXrqINBm9xWkcyUy53IQOj042dOs,22
 flexeval/__init__.py,sha256=UXI_xdSxnGAK2plDODBbPF3df-N7E9YJ418QHK7XN-Q,391
 flexeval/__main__.py,sha256=c9NQqsea3e-_6b736gBeIO3O_zdXQ1wtY3-Scj5NiPg,126
 flexeval/cli.py,sha256=RwtRk121OivbLQyYpYxJ7PugPIYQ8J4qXHFN2SxxPy4,2985
 flexeval/completions.py,sha256=pi_tYK4m3vKSqAC1ym9Jc3e4srcQSXfx-mX4qI5qisQ,5686
 flexeval/compute_metrics.py,sha256=4X6XFk0qUKcaCDllNeJreuhlnDHmfRPlsf0f8fWFOxA,37277
 flexeval/config.yaml,sha256=dpkFdW0rKf7StGoVeIGaCNw0n0yOfYWig0xmIfsDdbg,530
-flexeval/data_loader.py,sha256=EKc6wdpQuhrB2ai2U_fQxojzt1RR716ELisiZXpfu58,25311
+flexeval/data_loader.py,sha256=UP-HWqh5o_euqT2GvTbUYmA-yJcbTKtmug4w63w2CbA,26153
 flexeval/db_utils.py,sha256=2jgqexLCAqShvgPrImZz12UkMZtfERhP8iXjratXYok,1612
 flexeval/dependency_graph.py,sha256=SaG9gjkw2Q0NykqQWs4JzPkv5sMj2aXXmhjJ7yRkV4Q,10539
 flexeval/eval_schema.json,sha256=BQetj8O0_4rorj3Mpqk-sj_SCaRkGMrvBUcxhuw6zLE,13111
@@ -20,9 +20,10 @@ flexeval/classes/base.py,sha256=xxkTa8joPe39CFwveeTPW56LW-x7rsi5oBAIxrvM5iI,944
 flexeval/classes/dataset.py,sha256=Y_EdEIuhx526SSvkqk2tFBzkOgBkVY-5FeraYMtU5lo,2913
 flexeval/classes/eval_runner.py,sha256=ZvCpyaD7lorDK_mYJSZqQbvI6FfLbIWRFHNarWTAMQU,6270
 flexeval/classes/eval_set_run.py,sha256=fq_wBOaxuq7dLxiZIw76WGIwhRBNbQWDUhpiK0wDG_A,1116
-flexeval/classes/message.py,sha256=zuDm_v1gmK49Fw5m-HTWiqndrI_xtLotlXD8nhRDDTg,7518
+flexeval/classes/jsonview.py,sha256=3XJTh46ODfqdNbrXYDEV6kRO8KbeiHJo5pb4aJrbHRY,3459
+flexeval/classes/message.py,sha256=gDejDfaHGQKgS_CpJqjPAVzpiRD2JddKo17Yi1wVeiw,7676
 flexeval/classes/metric.py,sha256=d8l39_QwnQDmTJvy9TIulU4p0jqD7ldMUi4m5zfK2Es,2806
-flexeval/classes/thread.py,sha256=LchsK9mmrY4K-zSTMAAmywlzPVwnpZ7rOHqBGPIlda8,2779
+flexeval/classes/thread.py,sha256=cFQu3Mwzk8-Def8xccB8F6zKv64Srvhz5n83yLELvKo,2922
 flexeval/classes/tool_call.py,sha256=CteT2Hajor0PlHEEn7apfZux5_mremSIDrQmZ0iB7K0,1748
 flexeval/classes/turn.py,sha256=kLmgnYQ-4a8sydzGK1HTQRyUDXZIedmt_NFR3shLJFE,8635
 flexeval/configuration/__init__.py,sha256=wP_gpYyaEp5DxCSH8-4KHchH07JMZZOk8eCFMfd5LBw,75
@@ -41,8 +42,8 @@ flexeval/schema/eval_schema.py,sha256=iHMbanW4Ef_sp51KiaZKeP3Dn4Z6pWCGa7N2SPvsFK
 flexeval/schema/evalrun_schema.py,sha256=M7JY01DhlLzwZc2jJTIeGPs9vt6TFMPir51MFhtRllA,3526
 flexeval/schema/rubric_schema.py,sha256=uxcf7MHWKW3EmABUnWeCinGUP6LBjskiq7zkEPHmAvU,1615
 flexeval/schema/schema_utils.py,sha256=Fg1foqRA-9X-hl_vqIF3bpYdE51hNEgdw739Q-s3iQc,698
-python_flexeval-0.2.0.dist-info/METADATA,sha256=bEifn06Ok5-8YllS4uYxBN2KNuZvf7vJg8b_GarkttU,5599
-python_flexeval-0.2.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-python_flexeval-0.2.0.dist-info/entry_points.txt,sha256=wSyluqXhrX3xySVYAtM-Kv23p4OauKQCSBuNNfzEGtI,52
-python_flexeval-0.2.0.dist-info/licenses/LICENSE,sha256=OlAu_c13gw6-fJ9UdhZBMeNr5STLrnWG_0Hv0SCXtu4,1082
-python_flexeval-0.2.0.dist-info/RECORD,,
+python_flexeval-0.3.0.dist-info/METADATA,sha256=xBbeZrF4aEdl94pg-L2P_Di6cxtxA3aZnu6fxFjUf-8,5599
+python_flexeval-0.3.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+python_flexeval-0.3.0.dist-info/entry_points.txt,sha256=wSyluqXhrX3xySVYAtM-Kv23p4OauKQCSBuNNfzEGtI,52
+python_flexeval-0.3.0.dist-info/licenses/LICENSE,sha256=OlAu_c13gw6-fJ9UdhZBMeNr5STLrnWG_0Hv0SCXtu4,1082
+python_flexeval-0.3.0.dist-info/RECORD,,