qtype 0.1.3__tar.gz → 0.1.5__tar.gz

{qtype-0.1.3/qtype.egg-info → qtype-0.1.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: qtype
-Version: 0.1.3
+Version: 0.1.5
 Summary: DSL for Generative AI Prototyping
 Author-email: Lou Kratz <lou.kratz+qtype@bazaarvoice.com>
 License-Expression: Apache-2.0

{qtype-0.1.3 → qtype-0.1.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "qtype"
-version = "0.1.3"
+version = "0.1.5"
 description = "DSL for Generative AI Prototyping"
 authors = [{ name="Lou Kratz", email="lou.kratz+qtype@bazaarvoice.com" }]
 readme = "README.md"

{qtype-0.1.3 → qtype-0.1.5}/qtype/dsl/model.py

@@ -765,8 +765,9 @@ class AWSAuthProvider(AuthorizationProvider):
         has_keys = self.access_key_id and self.secret_access_key
         has_profile = self.profile_name
         has_role = self.role_arn
+        has_region = self.region
 
-        if not (has_keys or has_profile or has_role):
+        if not (has_keys or has_profile or has_role or has_region):
             raise ValueError(
                 "AWSAuthProvider must specify at least one authentication method: "
                 "access keys, profile name, or role ARN."

{qtype-0.1.3 → qtype-0.1.5}/qtype/dsl/parser.py

@@ -153,12 +153,12 @@ def _format_validation_errors(
         error_msg = "Validation failed (see details above)"
     else:
         error_msg = "Validation failed:\n"
-        for error in relevant_errors[:5]:  # Show max 5 errors
+        for error in relevant_errors[:30]:  # Show max 5 errors
             loc_path = _simplify_field_path(error["loc"])
             error_msg += f"  {loc_path}: {error['msg']}\n"
 
-        if len(relevant_errors) > 5:
-            error_msg += f"  ... and {len(relevant_errors) - 5} more errors\n"
+        if len(relevant_errors) > 30:
+            error_msg += f"  ... and {len(relevant_errors) - 30} more errors\n"
 
     if source_name:
         error_msg = f"In {source_name}:\n{error_msg}"

{qtype-0.1.3 → qtype-0.1.5}/qtype/interpreter/conversions.py

@@ -34,6 +34,7 @@ from qtype.interpreter.base.secrets import SecretManagerBase
 from qtype.interpreter.types import InterpreterError
 from qtype.semantic.model import (
     APIKeyAuthProvider,
+    AWSAuthProvider,
     DocumentIndex,
     DocumentSplitter,
     Model,
@@ -302,7 +303,9 @@ def to_vector_store(
 
 
 @cached_resource
-def to_embedding_model(model: Model) -> BaseEmbedding:
+def to_embedding_model(
+    model: Model, secret_manager: SecretManagerBase
+) -> BaseEmbedding:
     """Convert a qtype Model to a LlamaIndex embedding model."""
 
     if model.provider == "aws-bedrock":
@@ -310,7 +313,14 @@ def to_embedding_model(model: Model) -> BaseEmbedding:
             BedrockEmbedding,
         )
 
+        session = None
+        if model.auth is not None:
+            assert isinstance(model.auth, AWSAuthProvider)
+            with aws(model.auth, secret_manager) as session:
+                session = session._session
+
         bedrock_embedding: BaseEmbedding = BedrockEmbedding(
+            botocore_session=session,
             model_name=model.model_id if model.model_id else model.id,
             max_retries=100,
         )
@@ -320,8 +330,20 @@ def to_embedding_model(model: Model) -> BaseEmbedding:
             OpenAIEmbedding,
         )
 
+        api_key = None
+        if model.auth:
+            with auth(model.auth, secret_manager) as provider:
+                if not isinstance(provider, APIKeyAuthProvider):
+                    raise InterpreterError(
+                        f"OpenAI provider requires APIKeyAuthProvider, "
+                        f"got {type(provider).__name__}"
+                    )
+                # api_key is guaranteed to be str after auth() resolves it
+                api_key = provider.api_key  # type: ignore[assignment]
+
         openai_embedding: BaseEmbedding = OpenAIEmbedding(
-            model_name=model.model_id if model.model_id else model.id
+            api_key=api_key,
+            model_name=model.model_id if model.model_id else model.id,
         )
         return openai_embedding
     else:
@@ -566,7 +588,7 @@ def to_llama_vector_store_and_retriever(
     vector_store = to_vector_store(index, secret_manager)
 
     # Get the embedding model
-    embedding_model = to_embedding_model(index.embedding_model)
+    embedding_model = to_embedding_model(index.embedding_model, secret_manager)
 
     # Create a VectorStoreIndex with the vector store and embedding model
     vector_index = VectorStoreIndex.from_vector_store(

{qtype-0.1.3 → qtype-0.1.5}/qtype/interpreter/executors/document_embedder_executor.py

@@ -1,3 +1,4 @@
+import logging
 from typing import AsyncIterator
 
 from botocore.exceptions import ClientError
@@ -41,7 +42,7 @@ class DocumentEmbedderExecutor(StepExecutor):
         self.step: DocumentEmbedder = step
         # Initialize the embedding model once for the executor
         self.embedding_model: BaseEmbedding = to_embedding_model(
-            self.step.model
+            self.step.model, context.secret_manager
         )
 
     # TODO: properly abstract this into a mixin
@@ -103,5 +104,9 @@ class DocumentEmbedderExecutor(StepExecutor):
         except Exception as e:
             # Emit error event to stream so frontend can display it
             await self.stream_emitter.error(str(e))
+            logging.error(
+                f"Error processing DocumentEmbedder step {self.step.id}",
+                exc_info=e,
+            )
             message.set_error(self.step.id, e)
             yield message

{qtype-0.1.3 → qtype-0.1.5}/qtype/interpreter/executors/invoke_embedding_executor.py

@@ -30,7 +30,9 @@ class InvokeEmbeddingExecutor(StepExecutor):
             )
         self.step: InvokeEmbedding = step
         # Initialize the embedding model once for the executor
-        self.embedding_model = to_embedding_model(self.step.model)
+        self.embedding_model = to_embedding_model(
+            self.step.model, context.secret_manager
+        )
 
     async def process_message(
         self,

{qtype-0.1.3 → qtype-0.1.5/qtype.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: qtype
-Version: 0.1.3
+Version: 0.1.5
 Summary: DSL for Generative AI Prototyping
 Author-email: Lou Kratz <lou.kratz+qtype@bazaarvoice.com>
 License-Expression: Apache-2.0