openai-sdk-helpers 0.6.4__tar.gz → 0.6.6__tar.gz

{openai_sdk_helpers-0.6.4 → openai_sdk_helpers-0.6.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openai-sdk-helpers
-Version: 0.6.4
+Version: 0.6.6
 Summary: Composable helpers for OpenAI SDK agents, prompts, and storage
 Author: openai-sdk-helpers maintainers
 License: MIT

{openai_sdk_helpers-0.6.4 → openai_sdk_helpers-0.6.6}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "openai-sdk-helpers"
-version = "0.6.4"
+version = "0.6.6"
 requires-python = ">=3.10"
 readme = "README.md"
 description = "Composable helpers for OpenAI SDK agents, prompts, and storage"

{openai_sdk_helpers-0.6.4 → openai_sdk_helpers-0.6.6}/src/openai_sdk_helpers/agent/classifier.py

@@ -26,6 +26,8 @@ from ..utils import ensure_list
 from .base import AgentBase
 from .configuration import AgentConfiguration
 
+_CONTINUE_CONFIDENCE_THRESHOLD = 0.7
+
 
 class TaxonomyClassifierAgent(AgentBase):
     """Classify text by recursively traversing a taxonomy.
@@ -394,6 +396,13 @@ class TaxonomyClassifierAgent(AgentBase):
 
         resolved_nodes = _resolve_nodes(node_paths, step)
 
+        should_continue = _should_continue_from_stop(step, resolved_nodes)
+        if should_continue:
+            step = step.model_copy(
+                update={"stop_reason": ClassificationStopReason.CONTINUE}
+            )
+            state.steps[-1] = step
+
         if step.stop_reason.is_terminal:
             if resolved_nodes:
                 state.final_nodes.extend(resolved_nodes)
@@ -407,7 +416,7 @@ class TaxonomyClassifierAgent(AgentBase):
             return
 
         base_steps_len = len(state.steps)
-        child_tasks: list[tuple[Awaitable["_TraversalState"], int]] = []
+        child_tasks: list[tuple[Awaitable["_TraversalState"], int, TaxonomyNode]] = []
         for node in resolved_nodes:
             if node.children:
                 sub_agent = self._build_sub_agent(list(node.children))
@@ -429,6 +438,7 @@ class TaxonomyClassifierAgent(AgentBase):
                             state=sub_state,
                         ),
                         base_final_nodes_len,
+                        node,
                     )
                 )
             else:
@@ -439,13 +449,19 @@ class TaxonomyClassifierAgent(AgentBase):
                 )
         if child_tasks:
             child_states = await asyncio.gather(
-                *(child_task for child_task, _ in child_tasks)
+                *(child_task for child_task, _, _ in child_tasks)
             )
-            for child_state, (_, base_final_nodes_len) in zip(
+            for child_state, (_, base_final_nodes_len, parent_node) in zip(
                 child_states, child_tasks, strict=True
             ):
                 state.steps.extend(child_state.steps[base_steps_len:])
-                state.final_nodes.extend(child_state.final_nodes[base_final_nodes_len:])
+                child_final_nodes = child_state.final_nodes[base_final_nodes_len:]
+                state.final_nodes.extend(child_final_nodes)
+                if should_continue and not child_final_nodes:
+                    state.final_nodes.append(parent_node)
+                    state.best_confidence = _max_confidence(
+                        state.best_confidence, step.confidence
+                    )
                 state.best_confidence = _max_confidence(
                     state.best_confidence, child_state.best_confidence
                 )
@@ -627,6 +643,34 @@ def _resolve_stop_reason(state: _TraversalState) -> ClassificationStopReason:
     return ClassificationStopReason.NO_MATCH
 
 
+def _should_continue_from_stop(
+    step: ClassificationStep,
+    resolved_nodes: Sequence[TaxonomyNode],
+    threshold: float = _CONTINUE_CONFIDENCE_THRESHOLD,
+) -> bool:
+    """Return True when a stop reason should continue traversal.
+
+    Parameters
+    ----------
+    step : ClassificationStep
+        Classification step to evaluate.
+    resolved_nodes : Sequence[TaxonomyNode]
+        Resolved taxonomy nodes from the classification step.
+    threshold : float, default=0.7
+        Confidence threshold for overriding a stop reason.
+
+    Returns
+    -------
+    bool
+        True when traversal should proceed despite a stop reason.
+    """
+    if step.stop_reason is not ClassificationStopReason.STOP:
+        return False
+    if step.confidence is None or step.confidence < threshold:
+        return False
+    return any(node.children for node in resolved_nodes)
+
+
 def _normalize_roots(
     taxonomy: TaxonomyNode | Sequence[TaxonomyNode],
 ) -> list[TaxonomyNode]:

{openai_sdk_helpers-0.6.4 → openai_sdk_helpers-0.6.6}/src/openai_sdk_helpers/structure/base.py

@@ -174,6 +174,100 @@ def _ensure_schema_has_type(schema: dict[str, Any]) -> None:
     schema.update(_build_any_value_schema())
 
 
+def _strip_ref_types(
+    schema: dict[str, Any],
+    *,
+    nullable_fields: set[str] | None = None,
+) -> None:
+    """Remove type entries from enum $ref nodes when non-nullable.
+
+    Parameters
+    ----------
+    schema : dict[str, Any]
+        Root schema to clean in place.
+    nullable_fields : set[str] | None, optional
+        Field names that should remain nullable. Defaults to None.
+    """
+    field_names = nullable_fields or set()
+
+    def _resolve_ref(ref: str) -> dict[str, Any] | None:
+        if not ref.startswith("#/"):
+            return None
+        pointer = ref.removeprefix("#/")
+        current: Any = schema
+        for part in pointer.split("/"):
+            if not isinstance(current, dict):
+                return None
+            current = current.get(part)
+        if isinstance(current, dict):
+            return current
+        return None
+
+    def _is_enum_ref(ref: str) -> bool:
+        ref_target = _resolve_ref(ref)
+        if not isinstance(ref_target, dict):
+            return False
+        return "enum" in ref_target
+
+    def _is_null_schema(entry: Any) -> bool:
+        if not isinstance(entry, dict):
+            return False
+        entry_type = entry.get("type")
+        if entry_type == "null":
+            return True
+        if isinstance(entry_type, list) and "null" in entry_type:
+            return True
+        return False
+
+    def _walk(
+        node: Any,
+        *,
+        nullable_context: bool = False,
+        nullable_property: bool = False,
+    ) -> None:
+        if isinstance(node, dict):
+            ref = node.get("$ref")
+            if (
+                isinstance(ref, str)
+                and _is_enum_ref(ref)
+                and not nullable_context
+                and not nullable_property
+            ):
+                node.pop("type", None)
+            for key, value in node.items():
+                if key == "anyOf" and isinstance(value, list):
+                    anyof_nullable = any(_is_null_schema(entry) for entry in value)
+                    for entry in value:
+                        _walk(
+                            entry,
+                            nullable_context=anyof_nullable or nullable_context,
+                            nullable_property=nullable_property,
+                        )
+                    continue
+                if key == "properties" and isinstance(value, dict):
+                    for prop_name, prop_schema in value.items():
+                        _walk(
+                            prop_schema,
+                            nullable_context=nullable_context,
+                            nullable_property=prop_name in field_names,
+                        )
+                    continue
+                _walk(
+                    value,
+                    nullable_context=nullable_context,
+                    nullable_property=nullable_property,
+                )
+        elif isinstance(node, list):
+            for item in node:
+                _walk(
+                    item,
+                    nullable_context=nullable_context,
+                    nullable_property=nullable_property,
+                )
+
+    _walk(schema)
+
+
 def _hydrate_ref_types(schema: dict[str, Any]) -> None:
     """Attach explicit types to $ref nodes when available.
 
@@ -556,17 +650,18 @@ class StructureBase(BaseModelJSONSerializable):
 
         cleaned_schema = cast(dict[str, Any], clean_refs(schema))
 
-        cleaned_schema = cast(dict[str, Any], cleaned_schema)
-        _hydrate_ref_types(cleaned_schema)
-        _ensure_items_have_schema(cleaned_schema)
-        _ensure_schema_has_type(cleaned_schema)
-
         nullable_fields = {
             name
             for name, model_field in getattr(cls, "model_fields", {}).items()
             if getattr(model_field, "default", inspect.Signature.empty) is None
         }
 
+        cleaned_schema = cast(dict[str, Any], cleaned_schema)
+        _hydrate_ref_types(cleaned_schema)
+        _ensure_items_have_schema(cleaned_schema)
+        _ensure_schema_has_type(cleaned_schema)
+        _strip_ref_types(cleaned_schema, nullable_fields=nullable_fields)
+
         properties = cleaned_schema.get("properties", {})
         if isinstance(properties, dict) and nullable_fields:
             for field_name in nullable_fields:

{openai_sdk_helpers-0.6.4 → openai_sdk_helpers-0.6.6}/src/openai_sdk_helpers/structure/classification.py

@@ -480,6 +480,8 @@ class ClassificationResult(StructureBase):
         Yield selected identifiers across all steps.
     selected_nodes
         Return the selected identifiers across all steps.
+    to_lightweight_summary
+        Return a lightweight summary of selected node paths.
 
     Examples
     --------
@@ -565,6 +567,52 @@ class ClassificationResult(StructureBase):
                 if normalized:
                     yield normalized
 
+    def to_lightweight_summary(self) -> "ClassificationSummary | None":
+        """Return a lightweight summary of selected node paths.
+
+        Returns
+        -------
+        ClassificationSummary or None
+            Summary containing the ``full_paths`` of selected nodes, or None
+            when no selections exist.
+
+        Examples
+        --------
+        >>> result = ClassificationResult(steps=[])
+        >>> result.to_lightweight_summary() is None
+        True
+        """
+        full_paths = [
+            format_path_identifier(taxonomy_enum_path(node))
+            for node in self.selected_nodes
+        ]
+        if not full_paths:
+            return None
+        return ClassificationSummary(full_paths=full_paths)
+
+
+class ClassificationSummary(StructureBase):
+    """Represent a lightweight summary of selected taxonomy paths.
+
+    Attributes
+    ----------
+    full_paths : list[str] or None
+        Selected taxonomy paths including parent segments.
+
+    Methods
+    -------
+    to_json()
+        Serialize the summary to a JSON-compatible dictionary.
+    from_json(data)
+        Construct a summary instance from JSON-compatible data.
+    """
+
+    full_paths: list[str] | None = spec_field(
+        "full_paths",
+        description="Selected taxonomy paths including parent segments.",
+        default=None,
+    )
+
 
 def taxonomy_enum_path(value: Enum | str | None) -> list[str]:
     """Return the taxonomy path segments for an enum value.
@@ -598,6 +646,7 @@ def taxonomy_enum_path(value: Enum | str | None) -> list[str]:
 
 __all__ = [
     "ClassificationResult",
+    "ClassificationSummary",
     "ClassificationStep",
     "ClassificationStopReason",
     "Taxonomy",