mirror of
https://github.com/microsoft/semantic-kernel.git
synced 2026-10-12 02:19:06 +00:00
Eduard van Valkenburg
b91529a62e
### Motivation and Context <!-- Thank you for your contribution to the semantic-kernel repo! Please help reviewers and future users, providing the following information: 1. Why is this change required? 2. What problem does it solve? 3. What scenario does it contribute to? 4. If it fixes an open issue, please link to the issue here. --> This PR implements the new approach to AI Request settings as described here: https://github.com/microsoft/semantic-kernel/blob/main/docs/decisions/0008-support-generic-llm-request-settings.md and completes #3312 ### Description <!-- Describe your changes, the overall approach, the underlying design. These notes will help understanding how your code works. Thanks! --> Does this: - Creates AI Request Settings with service_id (referring to the registered AI services in a kernel) and extension_data, a dict that can hold any number of settings - has several methods: - `from_ai_request_settings` -> constructor to create a new request_settings object based on another, supports creating specific ai_request_settings from generic ones. - `update_from_ai_request_settings` -> updating the fields of a request setting from another ones. - `prepare_settings_dict` -> method that is used when actually creating the dict that get's passed to the handler of the AI service. - Created subclasses like OpenAIChatRequestSettings which has all the required fields of a OpenAI Chat completion, same for OpenAIText, AzureChat, GooglePalm, HuggingFace, etc. - added `get_request_settings_class` to ai_services to return the class used for the settings for that service. The goal is that when running, this settings object get's updated and a single operations (called `prepare_settings_dict`) is used to return a dict that can be passed to the api call directly, thereby reducing the need for custom logic everywhere, this will also negate the need for things like `complete_chat_with_functions_async` and others. So also removed complete_chat_..._async methods from openai and azure openai. It has also changed the prompt_template_config with the generic type for the AIRequestSettings or subclasses, and some other changes. ### Contribution Checklist <!-- Before submitting this PR, please make sure: --> - [x] The code builds clean without any errors or warnings - [x] The PR follows the [SK Contribution Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md) and the [pre-submission formatting script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts) raises no violations - [x] All unit tests pass, and I have added new tests where possible - [ ] I didn't break anyone 😄
27 lines
732 B
INI
27 lines
732 B
INI
[run]
|
|
source = semantic_kernel
|
|
omit =
|
|
semantic_kernel/connectors/memory/*
|
|
semantic_kernel/connectors/openapi/*
|
|
semantic_kernel/connectors/search_engine/*
|
|
semantic_kernel/connectors/ai/google_palm/*
|
|
semantic_kernel/connectors/ai/hugging_face/*
|
|
|
|
|
|
[report]
|
|
# Regexes for lines to exclude from consideration
|
|
exclude_lines =
|
|
# Have to re-enable the standard pragma
|
|
pragma: no cover
|
|
|
|
# Don't complain about missing debug-only code:
|
|
def __repr__
|
|
|
|
# Don't complain if tests don't hit defensive assertion code:
|
|
raise AssertionError
|
|
raise NotImplementedError
|
|
|
|
# TYPE_CHECKING and @overload blocks are never executed during pytest run
|
|
if TYPE_CHECKING:
|
|
@overload
|
|
@abstractmethod |