feat: streaming support in m serve OpenAI API server#823

Merged

markstur merged 23 commits intogenerative-computing:mainfrom

markstur:issue_822

Apr 17, 2026

Commits on Apr 16, 2026

feat: streaming support in m serve OpenAI API server
markstur
committed
test: add streaming test file
markstur
committed
fix: docstring fix for streaming
markstur
committed
fix: make system_fingerprint consistent in m serve
markstur
committed
fix: move stream_chat_completion_chunks out of helpers and into cli
markstur
committed
fix: astream() contract inconsistency in docstring and streaming mocks
markstur
committed
fix: emit [DONE] after error when streaming
markstur
committed
fix: do not add stream=False by default (regression)
markstur
committed
fix: remove misleading asyncio tags in test_serve_streaming.py
markstur
committed
fix: only include usage in stream results when requested
markstur
committed
fix: add missing code blocks in docs/examples/m_serve/README.md
markstur
committed
fix: add missing dict type details
markstur
committed
fix: m serve streaming initial and final chunks use content=None
markstur
committed
test: add test to confirm bug when mot is already computed
markstur
committed
fix: add logic before streaming loop to handle already computed chunk
markstur
committed
fix: remove unused import in openai_compatible_helpers.py
markstur
committed
fix: fix model comment about stream_options/incude_usage default behavior
markstur
committed
fix: typing fix
markstur
committed
fix: use pedantic model for m serve stream_options
markstur
committed
test: add validation unit tests for StreamOptions model
markstur
committed
fix: add logging of streaming exception
markstur
committed
fix: handle empty string content in streaming
markstur
committed
fix: FancyLogger is no longer Fancy
markstur
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: streaming support in m serve OpenAI API server#823

feat: streaming support in m serve OpenAI API server#823
markstur merged 23 commits intogenerative-computing:mainfrom
markstur:issue_822

Commits on Apr 16, 2026

feat: streaming support in m serve OpenAI API server

test: add streaming test file

fix: docstring fix for streaming

fix: make system_fingerprint consistent in m serve

fix: move stream_chat_completion_chunks out of helpers and into cli

fix: astream() contract inconsistency in docstring and streaming mocks

fix: emit [DONE] after error when streaming

fix: do not add stream=False by default (regression)

fix: remove misleading asyncio tags in test_serve_streaming.py

fix: only include usage in stream results when requested

fix: add missing code blocks in docs/examples/m_serve/README.md

fix: add missing dict type details

fix: m serve streaming initial and final chunks use content=None

test: add test to confirm bug when mot is already computed

fix: add logic before streaming loop to handle already computed chunk

fix: remove unused import in openai_compatible_helpers.py

fix: fix model comment about stream_options/incude_usage default behavior

fix: typing fix

fix: use pedantic model for m serve stream_options

test: add validation unit tests for StreamOptions model

fix: add logging of streaming exception

fix: handle empty string content in streaming

fix: FancyLogger is no longer Fancy