feat(sdk/python): add Responses API client by MaanavD · Pull Request #670 · microsoft/Foundry-Local

MaanavD · 2026-04-23T16:45:53Z

Summary

Implements the OpenAI Responses API client for the Foundry Local Python SDK. HTTP-only, sync pattern matching the existing chat_client.py style.

New files

File	Description
`src/openai/responses_types.py`	Full type system: content parts, response items, tools, config, `ResponseObject` (with `output_text` property), all streaming event dataclasses, `parse_streaming_event` factory, `_to_dict` serializer
`src/openai/responses_client.py`	HTTP client: `ResponsesClient`, `ResponsesClientSettings`, `ResponsesAPIError`. Methods: `create`, `create_streaming` (SSE generator), `get`, `delete`, `cancel`, `get_input_items`, `list`
`examples/responses.py`	5 end-to-end scenarios: basic create, streaming, multi-turn, tool calling, vision
`test/openai/test_responses_client.py`	56 unit tests with mocked HTTP
`test/openai/test_responses_integration.py`	14 integration tests (gated on `FOUNDRY_INTEGRATION_TESTS=1`)

Modified files

foundry_local_manager.py — create_responses_client(model_id) factory
imodel.py / detail/model.py / detail/model_variant.py — factory wired through the model hierarchy
src/__init__.py / src/openai/__init__.py — all new public types exported

Test results

Unit tests: 56/56 passing (no server needed)
Integration tests: 14/14 passing against live qwen2.5-0.5b server

Implements the OpenAI Responses API client for the Foundry Local Python SDK. New files: - src/openai/responses_types.py: full type system (content parts, items, tools, config, ResponseObject with output_text property), all streaming event dataclasses, parse_streaming_event factory, and _to_dict serializer - src/openai/responses_client.py: HTTP-only sync client (ResponsesClient, ResponsesClientSettings, ResponsesAPIError) with create, create_streaming (SSE generator), get, delete, cancel, get_input_items, list - examples/responses.py: 5 end-to-end scenarios (basic, streaming, multi-turn, tool calling, vision) - test/openai/test_responses_client.py: 56 unit tests (mocked HTTP) - test/openai/test_responses_integration.py: 14 integration tests gated on FOUNDRY_INTEGRATION_TESTS=1 Modified files: - src/foundry_local_manager.py: create_responses_client factory method - src/imodel.py: abstract create_responses_client - src/detail/model.py: delegating create_responses_client - src/detail/model_variant.py: concrete create_responses_client - src/openai/__init__.py: export ResponsesClient, ResponsesClientSettings, ResponsesAPIError - src/__init__.py: export all public Responses API types Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

vercel · 2026-04-23T16:45:58Z

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
foundry-local	Ready	Preview, Comment	May 1, 2026 9:57pm

Copilot

Pull request overview

Adds an OpenAI-compatible Responses API surface to the Foundry Local Python SDK, implemented as an HTTP client against the embedded web service and wired through the manager/model abstractions.

Changes:

Introduces ResponsesClient (+ settings + error type) with non-streaming and SSE-streaming support.
Adds a full Responses DTO/type system (content parts, items, tools/config, response object helpers, streaming event parsing/serialization).
Adds examples and both mocked unit tests and gated live integration tests; wires factory methods through FoundryLocalManager and IModel implementations.

Reviewed changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
sdk/python/src/openai/responses_client.py	New HTTP client for `/v1/responses`, including SSE parsing helpers.
sdk/python/src/openai/responses_types.py	New dataclass-based type system + parsers/serializers for Responses API.
sdk/python/src/openai/init.py	Exports Responses client types from `foundry_local_sdk.openai`.
sdk/python/src/foundry_local_manager.py	Adds `create_responses_client()` factory bound to running web service.
sdk/python/src/imodel.py	Extends `IModel` with `create_responses_client(base_url)` API.
sdk/python/src/detail/model.py	Wires `Model.create_responses_client()` through the selected variant.
sdk/python/src/detail/model_variant.py	Implements `create_responses_client()` for a specific variant.
sdk/python/src/init.py	Exposes Responses client + public DTOs at package root.
sdk/python/examples/responses.py	End-to-end example covering create, streaming, tool calls, vision.
sdk/python/test/openai/test_responses_client.py	Mocked unit tests for request building, SSE parsing, error handling, serialization.
sdk/python/test/openai/test_responses_integration.py	Gated live integration tests against a real runtime/web service.
sdk/python/requirements.txt	Updates pinned native/runtime dependencies and platform markers.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

- Add configurable timeout to ResponsesClientSettings (default 60s); non-streaming calls use it directly, streaming uses it as connect timeout with unbounded read (suitable for long responses) - Fix SSE buffer: replace O(n) list-join-per-chunk with a single string buffer and split on double-newline; use chunk_size=None for natural server chunk boundaries - Add InputImageContent.__post_init__ to enforce exactly one of image_url or image_data (raises ValueError if both or neither) - Add optional max_size=(w,h) to InputImageContent.from_file and from_bytes to resize images before base64-encoding (requires Pillow) - Raise ValueError for unknown content-part types instead of silently returning a fallback InputTextContent - Document _MAX_ID_LEN=256 with rationale; lower from 1024 Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 12 out of 12 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

…urns None, example usage guard - ResponsesClientSettings.store defaults to None (omitted from request body, server decides) — aligns with JS SDK which has store?: boolean - _MAX_ID_LEN reverted to 1024 to align with JS SDK constant - _parse_content_part returns None for unknown types (forward-compat, not ValueError); _parse_content filters out None entries - examples/responses.py: guard event.response.usage chain with getattr to avoid AttributeError if response or usage is absent - Tests updated: store default tests, too-long-id threshold (1025), request body assertions Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…onses-api-python-sdk

- Rename model and variant Responses client factories to get_responses_client to match existing get_*_client naming. - Use FoundryLocalException for Responses API transport and parsing errors instead of exporting a dedicated ResponsesAPIError. - Keep only the foundry-local-core version bump in requirements.txt and restore existing ORT dependency markers/order. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Replace the SDK-native Responses client implementation with a focused Python sample and integration tests that use FoundryLocalManager for setup/model/server lifecycle and the official OpenAI Python client for /v1/responses calls. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…onses-api-python-sdk

MaanavD · 2026-05-01T20:09:31Z

Pivoted this PR to the web-service sample/test approach:

Removed the SDK-native Python Responses client/types and manager/model Responses APIs from this branch.
Added sdk/python/examples/responses_web_service.py, which uses FoundryLocalManager for setup/model/server lifecycle and the official OpenAI Python client for /v1/responses.
Added sdk/python/test/openai/test_responses_web_service.py covering non-streaming, streaming, and function/tool-call round trip via the local web service.
Updated the Python README examples list.
Merged latest origin/main so the PR diff is focused on the sample/test changes.

Local validation: python -m pytest test\openai\test_responses_web_service.py --no-header -q => 3 passed (with the existing pytest timeout config warning).

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot AI review requested due to automatic review settings April 23, 2026 16:45

MaanavD mentioned this pull request Apr 23, 2026

feat(sdk/cs): add Responses API client for OpenAI/OpenResponses compat #505

Closed

vercel Bot deployed to Preview April 23, 2026 16:46 View deployment

Copilot started reviewing on behalf of MaanavD April 23, 2026 16:46 View session

Copilot AI reviewed Apr 23, 2026

View reviewed changes

Comment thread sdk/python/src/openai/responses_client.py Outdated

Comment thread sdk/python/src/openai/responses_client.py Outdated

Comment thread sdk/python/src/openai/responses_types.py Outdated

apsonawane reviewed Apr 23, 2026

View reviewed changes

Comment thread sdk/python/src/openai/responses_types.py Outdated

apsonawane reviewed Apr 23, 2026

View reviewed changes

Comment thread sdk/python/src/openai/responses_client.py Outdated

apsonawane reviewed Apr 23, 2026

View reviewed changes

Comment thread sdk/python/src/openai/responses_client.py Outdated

apsonawane reviewed Apr 23, 2026

View reviewed changes

Comment thread sdk/python/src/openai/responses_types.py Outdated

vercel Bot deployed to Preview April 23, 2026 21:11 View deployment

apsonawane requested a review from Copilot April 24, 2026 00:00

Copilot started reviewing on behalf of apsonawane April 24, 2026 00:01 View session

Copilot AI reviewed Apr 24, 2026

View reviewed changes

Comment thread sdk/python/src/openai/responses_client.py Outdated

Comment thread sdk/python/src/openai/responses_client.py Outdated

Comment thread sdk/python/examples/responses.py Outdated

apsonawane reviewed Apr 24, 2026

View reviewed changes

Comment thread sdk/python/src/openai/responses_types.py Outdated

vercel Bot deployed to Preview April 24, 2026 02:04 View deployment

Merge remote-tracking branch 'origin/main' into agents/implement-resp…

a73cf7a

…onses-api-python-sdk

vercel Bot deployed to Preview April 27, 2026 17:15 View deployment

prathikr reviewed Apr 27, 2026

View reviewed changes

Comment thread sdk/python/src/detail/model.py Outdated

Comment thread sdk/python/src/detail/model_variant.py Outdated

Comment thread sdk/python/src/openai/__init__.py Outdated

Comment thread sdk/python/requirements.txt Outdated

vercel Bot deployed to Preview April 27, 2026 19:12 View deployment

MaanavD mentioned this pull request Apr 28, 2026

Add Responses web-service sample + integration tests for C# SDK #681

Draft

vercel Bot deployed to Preview May 1, 2026 20:07 View deployment

Merge remote-tracking branch 'origin/main' into agents/implement-resp…

6678462

…onses-api-python-sdk

vercel Bot deployed to Preview May 1, 2026 20:10 View deployment

refactor(sdk/python): align responses web-service sample tests

718e78c

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

vercel Bot deployed to Preview May 1, 2026 20:18 View deployment

docs(samples): add Python responses web-service README

0808187

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

vercel Bot deployed to Preview May 1, 2026 21:57 View deployment

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(sdk/python): add Responses API client#670

feat(sdk/python): add Responses API client#670
MaanavD wants to merge 9 commits intomainfrom
agents/implement-responses-api-python-sdk

MaanavD commented Apr 23, 2026

Uh oh!

vercel Bot commented Apr 23, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

MaanavD commented May 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

MaanavD commented Apr 23, 2026

Summary

New files

Modified files

Test results

Related

Uh oh!

vercel Bot commented Apr 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

MaanavD commented May 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

vercel Bot commented Apr 23, 2026 •

edited

Loading