openaivec 0.14.10__tar.gz → 0.14.12__tar.gz

{openaivec-0.14.10 → openaivec-0.14.12}/.github/copilot-instructions.md

@@ -24,7 +24,10 @@ Entry points:
 - Spark UDF builders in `spark.py`
 - Structured tasks under `task/`
 
-Azure note: Use deployment name as `model`. Warn if base URL not v1. Behavior otherwise mirrors OpenAI.
+Azure note: Use deployment name as `model`. Standard Azure OpenAI configuration uses:
+- Base URL: `https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/`
+- API Version: `"preview"`
+Warn if base URL not v1. Behavior otherwise mirrors OpenAI.
 
 ---
 
@@ -137,7 +140,16 @@ Public exports (`__init__.py`): `BatchResponses`, `AsyncBatchResponses`, `BatchE
 ## 10. Provider / Azure Rules
 
 - Auto-detect provider from env variables; deployment name = model for Azure.
-- Warn (don’t fail) if Azure base URL not v1 format; still proceed.
+- Standard Azure OpenAI configuration:
+  - Base URL: `https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/`
+  - API Version: `"preview"`
+  - Environment variables:
+    ```bash
+    export AZURE_OPENAI_API_KEY="your-azure-key"
+    export AZURE_OPENAI_BASE_URL="https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/"
+    export AZURE_OPENAI_API_VERSION="preview"
+    ```
+- Warn (don't fail) if Azure base URL not v1 format; still proceed.
 - Keep code paths unified; avoid forking logic unless behavior diverges.
 
 ---
@@ -348,6 +360,9 @@ uv run mkdocs serve
 Environment setup notes:
 
 - Set `OPENAI_API_KEY` or Azure trio (`AZURE_OPENAI_API_KEY`, `AZURE_OPENAI_BASE_URL`, `AZURE_OPENAI_API_VERSION`).
+- Standard Azure OpenAI configuration:
+  - `AZURE_OPENAI_BASE_URL="https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/"`
+  - `AZURE_OPENAI_API_VERSION="preview"`
 - Tests auto-skip live paths when credentials absent.
 - Use separate shell profiles per provider if switching frequently.
-- Azure canonical base URL should end with `/openai/v1/` (e.g. `https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/`); non‑v1 forms emit a warning.
+- Azure canonical base URL must end with `/openai/v1/` (e.g. `https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/`); non‑v1 forms emit a warning.

{openaivec-0.14.10 → openaivec-0.14.12}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openaivec
-Version: 0.14.10
+Version: 0.14.12
 Summary: Generative mutation for tabular calculation
 Project-URL: Homepage, https://microsoft.github.io/openaivec/
 Project-URL: Repository, https://github.com/microsoft/openaivec
@@ -334,26 +334,34 @@ Scale to enterprise datasets with distributed processing:
 First, obtain a Spark session and configure authentication:
 
 ```python
-import os
 from pyspark.sql import SparkSession
+from openaivec.spark import setup, setup_azure
 
 spark = SparkSession.builder.getOrCreate()
-sc = spark.sparkContext
 
-# Configure authentication via SparkContext environment variables
 # Option 1: Using OpenAI
-sc.environment["OPENAI_API_KEY"] = os.environ.get("OPENAI_API_KEY")
+setup(
+    spark,
+    api_key="your-openai-api-key",
+    responses_model_name="gpt-4.1-mini",  # Optional: set default model
+    embeddings_model_name="text-embedding-3-small"  # Optional: set default model
+)
 
 # Option 2: Using Azure OpenAI
-# sc.environment["AZURE_OPENAI_API_KEY"] = os.environ.get("AZURE_OPENAI_API_KEY")
-# sc.environment["AZURE_OPENAI_BASE_URL"] = os.environ.get("AZURE_OPENAI_BASE_URL")
-# sc.environment["AZURE_OPENAI_API_VERSION"] = os.environ.get("AZURE_OPENAI_API_VERSION")
+# setup_azure(
+#     spark,
+#     api_key="your-azure-openai-api-key",
+#     base_url="https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/",
+#     api_version="preview",
+#     responses_model_name="my-gpt4-deployment",  # Optional: set default deployment
+#     embeddings_model_name="my-embedding-deployment"  # Optional: set default deployment
+# )
 ```
 
 Next, create and register UDFs using the provided functions:
 
 ```python
-from openaivec.spark import responses_udf, task_udf, embeddings_udf, count_tokens_udf
+from openaivec.spark import responses_udf, task_udf, embeddings_udf, count_tokens_udf, similarity_udf, parse_udf
 from pydantic import BaseModel
 
 # --- Register Responses UDF (String Output) ---
@@ -387,6 +395,9 @@ spark.udf.register(
 # --- Register Token Counting UDF ---
 spark.udf.register("count_tokens", count_tokens_udf())
 
+# --- Register Similarity UDF ---
+spark.udf.register("compute_similarity", similarity_udf())
+
 # --- Register UDFs with Pre-configured Tasks ---
 from openaivec.task import nlp, customer_support
 
@@ -414,6 +425,17 @@ spark.udf.register(
     )
 )
 
+# --- Register Parse UDF (Dynamic Schema Inference) ---
+spark.udf.register(
+    "parse_dynamic",
+    parse_udf(
+        instructions="Extract key entities and attributes from the text",
+        example_table_name="sample_texts",  # Infer schema from examples
+        example_field_name="text",
+        max_examples=50
+    )
+)
+
 ```
 
 You can now use these UDFs in Spark SQL:
@@ -691,17 +713,19 @@ steps:
    - In the notebook, import and use `openaivec.spark` functions as you normally would. For example:
 
      ```python
-     import os
-     from openaivec.spark import responses_udf, embeddings_udf
+     from openaivec.spark import setup_azure, responses_udf, embeddings_udf
 
      # In Microsoft Fabric, spark session is automatically available
      # spark = SparkSession.builder.getOrCreate()
-     sc = spark.sparkContext
-
+     
      # Configure Azure OpenAI authentication
-     sc.environment["AZURE_OPENAI_API_KEY"] = "<your-api-key>"
-     sc.environment["AZURE_OPENAI_BASE_URL"] = "https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/"
-     sc.environment["AZURE_OPENAI_API_VERSION"] = "preview"
+     setup_azure(
+         spark,
+         api_key="<your-api-key>",
+         base_url="https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/",
+         api_version="preview",
+         responses_model_name="my-gpt4-deployment"  # Your Azure deployment name
+     )
 
      # Register UDFs
      spark.udf.register(

{openaivec-0.14.10 → openaivec-0.14.12}/README.md

@@ -308,26 +308,34 @@ Scale to enterprise datasets with distributed processing:
 First, obtain a Spark session and configure authentication:
 
 ```python
-import os
 from pyspark.sql import SparkSession
+from openaivec.spark import setup, setup_azure
 
 spark = SparkSession.builder.getOrCreate()
-sc = spark.sparkContext
 
-# Configure authentication via SparkContext environment variables
 # Option 1: Using OpenAI
-sc.environment["OPENAI_API_KEY"] = os.environ.get("OPENAI_API_KEY")
+setup(
+    spark,
+    api_key="your-openai-api-key",
+    responses_model_name="gpt-4.1-mini",  # Optional: set default model
+    embeddings_model_name="text-embedding-3-small"  # Optional: set default model
+)
 
 # Option 2: Using Azure OpenAI
-# sc.environment["AZURE_OPENAI_API_KEY"] = os.environ.get("AZURE_OPENAI_API_KEY")
-# sc.environment["AZURE_OPENAI_BASE_URL"] = os.environ.get("AZURE_OPENAI_BASE_URL")
-# sc.environment["AZURE_OPENAI_API_VERSION"] = os.environ.get("AZURE_OPENAI_API_VERSION")
+# setup_azure(
+#     spark,
+#     api_key="your-azure-openai-api-key",
+#     base_url="https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/",
+#     api_version="preview",
+#     responses_model_name="my-gpt4-deployment",  # Optional: set default deployment
+#     embeddings_model_name="my-embedding-deployment"  # Optional: set default deployment
+# )
 ```
 
 Next, create and register UDFs using the provided functions:
 
 ```python
-from openaivec.spark import responses_udf, task_udf, embeddings_udf, count_tokens_udf
+from openaivec.spark import responses_udf, task_udf, embeddings_udf, count_tokens_udf, similarity_udf, parse_udf
 from pydantic import BaseModel
 
 # --- Register Responses UDF (String Output) ---
@@ -361,6 +369,9 @@ spark.udf.register(
 # --- Register Token Counting UDF ---
 spark.udf.register("count_tokens", count_tokens_udf())
 
+# --- Register Similarity UDF ---
+spark.udf.register("compute_similarity", similarity_udf())
+
 # --- Register UDFs with Pre-configured Tasks ---
 from openaivec.task import nlp, customer_support
 
@@ -388,6 +399,17 @@ spark.udf.register(
     )
 )
 
+# --- Register Parse UDF (Dynamic Schema Inference) ---
+spark.udf.register(
+    "parse_dynamic",
+    parse_udf(
+        instructions="Extract key entities and attributes from the text",
+        example_table_name="sample_texts",  # Infer schema from examples
+        example_field_name="text",
+        max_examples=50
+    )
+)
+
 ```
 
 You can now use these UDFs in Spark SQL:
@@ -665,17 +687,19 @@ steps:
    - In the notebook, import and use `openaivec.spark` functions as you normally would. For example:
 
      ```python
-     import os
-     from openaivec.spark import responses_udf, embeddings_udf
+     from openaivec.spark import setup_azure, responses_udf, embeddings_udf
 
      # In Microsoft Fabric, spark session is automatically available
      # spark = SparkSession.builder.getOrCreate()
-     sc = spark.sparkContext
-
+     
      # Configure Azure OpenAI authentication
-     sc.environment["AZURE_OPENAI_API_KEY"] = "<your-api-key>"
-     sc.environment["AZURE_OPENAI_BASE_URL"] = "https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/"
-     sc.environment["AZURE_OPENAI_API_VERSION"] = "preview"
+     setup_azure(
+         spark,
+         api_key="<your-api-key>",
+         base_url="https://YOUR-RESOURCE-NAME.services.ai.azure.com/openai/v1/",
+         api_version="preview",
+         responses_model_name="my-gpt4-deployment"  # Your Azure deployment name
+     )
 
      # Register UDFs
      spark.udf.register(

{openaivec-0.14.10 → openaivec-0.14.12}/src/openaivec/_di.py

@@ -303,3 +303,24 @@ class Container:
             self._providers.clear()
             self._instances.clear()
             self._resolving.clear()
+
+    def clear_singletons(self) -> None:
+        """Clear all cached singleton instances from the container.
+
+        Removes all cached singleton instances while keeping the registered
+        providers intact. After calling this method, the next resolve call
+        for any service will create a new instance using the provider function.
+
+        Example:
+            ```python
+            container = Container()
+            container.register(str, lambda: "Hello")
+            instance1 = container.resolve(str)
+            container.clear_singletons()
+            instance2 = container.resolve(str)
+            print(instance1 is instance2)
+            # False - different instances after clearing singletons
+            ```
+        """
+        with self._lock:
+            self._instances.clear()

{openaivec-0.14.10 → openaivec-0.14.12}/src/openaivec/_provider.py

@@ -130,35 +130,9 @@ def provide_async_openai_client() -> AsyncOpenAI:
     )
 
 
-CONTAINER.register(ResponsesModelName, lambda: ResponsesModelName("gpt-4.1-mini"))
-CONTAINER.register(EmbeddingsModelName, lambda: EmbeddingsModelName("text-embedding-3-small"))
-CONTAINER.register(OpenAIAPIKey, lambda: OpenAIAPIKey(os.getenv("OPENAI_API_KEY")))
-CONTAINER.register(AzureOpenAIAPIKey, lambda: AzureOpenAIAPIKey(os.getenv("AZURE_OPENAI_API_KEY")))
-CONTAINER.register(AzureOpenAIBaseURL, lambda: AzureOpenAIBaseURL(os.getenv("AZURE_OPENAI_BASE_URL")))
-CONTAINER.register(
-    cls=AzureOpenAIAPIVersion,
-    provider=lambda: AzureOpenAIAPIVersion(os.getenv("AZURE_OPENAI_API_VERSION", "preview")),
-)
-CONTAINER.register(OpenAI, provide_openai_client)
-CONTAINER.register(AsyncOpenAI, provide_async_openai_client)
-CONTAINER.register(tiktoken.Encoding, lambda: tiktoken.get_encoding("o200k_base"))
-CONTAINER.register(TextChunker, lambda: TextChunker(CONTAINER.resolve(tiktoken.Encoding)))
-CONTAINER.register(
-    SchemaInferer,
-    lambda: SchemaInferer(
-        client=CONTAINER.resolve(OpenAI),
-        model_name=CONTAINER.resolve(ResponsesModelName).value,
-    ),
-)
-
-
-def reset_environment_registrations():
-    """Reset environment variable related registrations in the container.
-
-    This function re-registers environment variable dependent services to pick up
-    current environment variable values. Useful for testing when environment
-    variables are changed after initial container setup.
-    """
+def set_default_registrations():
+    CONTAINER.register(ResponsesModelName, lambda: ResponsesModelName("gpt-4.1-mini"))
+    CONTAINER.register(EmbeddingsModelName, lambda: EmbeddingsModelName("text-embedding-3-small"))
     CONTAINER.register(OpenAIAPIKey, lambda: OpenAIAPIKey(os.getenv("OPENAI_API_KEY")))
     CONTAINER.register(AzureOpenAIAPIKey, lambda: AzureOpenAIAPIKey(os.getenv("AZURE_OPENAI_API_KEY")))
     CONTAINER.register(AzureOpenAIBaseURL, lambda: AzureOpenAIBaseURL(os.getenv("AZURE_OPENAI_BASE_URL")))
@@ -168,6 +142,8 @@ def reset_environment_registrations():
     )
     CONTAINER.register(OpenAI, provide_openai_client)
     CONTAINER.register(AsyncOpenAI, provide_async_openai_client)
+    CONTAINER.register(tiktoken.Encoding, lambda: tiktoken.get_encoding("o200k_base"))
+    CONTAINER.register(TextChunker, lambda: TextChunker(CONTAINER.resolve(tiktoken.Encoding)))
     CONTAINER.register(
         SchemaInferer,
         lambda: SchemaInferer(
@@ -175,3 +151,6 @@ def reset_environment_registrations():
             model_name=CONTAINER.resolve(ResponsesModelName).value,
         ),
     )
+
+
+set_default_registrations()

{openaivec-0.14.10 → openaivec-0.14.12}/src/openaivec/pandas_ext.py

@@ -454,6 +454,7 @@ class OpenAIVecSeriesAccessor:
         """Parse Series values using an LLM with a provided cache.
         This method allows you to parse the Series content into structured data
         using an LLM, optionally inferring a schema based on the provided purpose.
+
         Args:
             instructions (str): System prompt for the LLM.
             cache (BatchingMapProxy[str, BaseModel]): Explicit cache instance for