release: 1.82.0 (#2372)

* Add background streaming

* -m rest of the implementation

* docs(readme): fix async example

* docs(readme): another async example fix

* fix(azure): mark images/edits as a deployment endpoint #2371

* feat(api): new streaming helpers for background responses

* release: 1.82.0

---------

Co-authored-by: pakrym-oai <[email protected]>
Co-authored-by: Robert Craigie <[email protected]>
Co-authored-by: Kevin Whinnery <[email protected]>
Co-authored-by: Friedel van Megen <[email protected]>
Co-authored-by: stainless-app[bot] <142633134+stainless-app[bot]@users.noreply.github.com>

Author: 5 people

.release-please-manifest.json

@@ -1,3 +1,3 @@
 {
-  ".": "1.81.0"
+  ".": "1.82.0"
 }

.stats.yml

@@ -1,4 +1,4 @@
 configured_endpoints: 111
-openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai%2Fopenai-6af14840a810139bf407013167ce1c8fb21b6ef8eb0cc3db58b51af7d52c4b5a.yml
-openapi_spec_hash: 3241bde6b273cfec0035e522bd07985d
-config_hash: 7367b68a4e7db36885c1a886f57b17f6
+openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai%2Fopenai-fc64d7c2c8f51f750813375356c3f3fdfc7fc1b1b34f19c20a5410279d445d37.yml
+openapi_spec_hash: 618285fc70199ee32b9ebe4bf72f7e4c
+config_hash: c497f6b750cc89c0bf2eefc0bc839c70

CHANGELOG.md

@@ -1,5 +1,24 @@
 # Changelog
 
+## 1.82.0 (2025-05-22)
+
+Full Changelog: [v1.81.0...v1.82.0](https://github.com/openai/openai-python/compare/v1.81.0...v1.82.0)
+
+### Features
+
+* **api:** new streaming helpers for background responses ([2a65d4d](https://github.com/openai/openai-python/commit/2a65d4de0aaba7801edd0df10f225530fd4969bd))
+
+
+### Bug Fixes
+
+* **azure:** mark images/edits as a deployment endpoint [#2371](https://github.com/openai/openai-python/issues/2371) ([5d1d5b4](https://github.com/openai/openai-python/commit/5d1d5b4b6072afe9fd7909b1a36014c8c11c1ad6))
+
+
+### Documentation
+
+* **readme:** another async example fix ([9ec8289](https://github.com/openai/openai-python/commit/9ec8289041f395805c67efd97847480f84eb9dac))
+* **readme:** fix async example ([37d0b25](https://github.com/openai/openai-python/commit/37d0b25b6e82cd381e5d1aa6e28f1a1311d02353))
+
 ## 1.81.0 (2025-05-21)
 
 Full Changelog: [v1.80.0...v1.81.0](https://github.com/openai/openai-python/compare/v1.80.0...v1.81.0)

README.md

@@ -174,13 +174,13 @@ client = AsyncOpenAI()
 
 
 async def main():
-    stream = client.responses.create(
+    stream = await client.responses.create(
         model="gpt-4o",
         input="Write a one-sentence bedtime story about a unicorn.",
         stream=True,
     )
 
-    for event in stream:
+    async for event in stream:
         print(event)
 
 

api.md

@@ -764,7 +764,6 @@ from openai.types.responses import (
     ResponseRefusalDoneEvent,
     ResponseStatus,
     ResponseStreamEvent,
-    ResponseTextAnnotationDeltaEvent,
     ResponseTextConfig,
     ResponseTextDeltaEvent,
     ResponseTextDoneEvent,

examples/responses/background.py

@@ -0,0 +1,46 @@
+from typing import List
+
+import rich
+from pydantic import BaseModel
+
+from openai import OpenAI
+
+
+class Step(BaseModel):
+    explanation: str
+    output: str
+
+
+class MathResponse(BaseModel):
+    steps: List[Step]
+    final_answer: str
+
+
+client = OpenAI()
+id = None
+
+with client.responses.create(
+    input="solve 8x + 31 = 2",
+    model="gpt-4o-2024-08-06",
+    background=True,
+    stream=True,
+) as stream:
+    for event in stream:
+        if event.type == "response.created":
+            id = event.response.id
+        if "output_text" in event.type:
+            rich.print(event)
+        if event.sequence_number == 10:
+            break
+
+print("Interrupted. Continuing...")
+
+assert id is not None
+with client.responses.retrieve(
+    response_id=id,
+    stream=True,
+    starting_after=10,
+) as stream:
+    for event in stream:
+        if "output_text" in event.type:
+            rich.print(event)

examples/responses/background_async.py

@@ -0,0 +1,52 @@
+import asyncio
+from typing import List
+
+import rich
+from pydantic import BaseModel
+
+from openai._client import AsyncOpenAI
+
+
+class Step(BaseModel):
+    explanation: str
+    output: str
+
+
+class MathResponse(BaseModel):
+    steps: List[Step]
+    final_answer: str
+
+
+async def main() -> None:
+    client = AsyncOpenAI()
+    id = None
+
+    async with await client.responses.create(
+        input="solve 8x + 31 = 2",
+        model="gpt-4o-2024-08-06",
+        background=True,
+        stream=True,
+    ) as stream:
+        async for event in stream:
+            if event.type == "response.created":
+                id = event.response.id
+            if "output_text" in event.type:
+                rich.print(event)
+            if event.sequence_number == 10:
+                break
+
+    print("Interrupted. Continuing...")
+
+    assert id is not None
+    async with await client.responses.retrieve(
+        response_id=id,
+        stream=True,
+        starting_after=10,
+    ) as stream:
+        async for event in stream:
+            if "output_text" in event.type:
+                rich.print(event)
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

examples/responses/background_streaming.py

@@ -0,0 +1,48 @@
+#!/usr/bin/env -S rye run python
+from typing import List
+
+import rich
+from pydantic import BaseModel
+
+from openai import OpenAI
+
+
+class Step(BaseModel):
+    explanation: str
+    output: str
+
+
+class MathResponse(BaseModel):
+    steps: List[Step]
+    final_answer: str
+
+
+client = OpenAI()
+id = None
+with client.responses.stream(
+    input="solve 8x + 31 = 2",
+    model="gpt-4o-2024-08-06",
+    text_format=MathResponse,
+    background=True,
+) as stream:
+    for event in stream:
+        if event.type == "response.created":
+            id = event.response.id
+        if "output_text" in event.type:
+            rich.print(event)
+        if event.sequence_number == 10:
+            break
+
+print("Interrupted. Continuing...")
+
+assert id is not None
+with client.responses.stream(
+    response_id=id,
+    starting_after=10,
+    text_format=MathResponse,
+) as stream:
+    for event in stream:
+        if "output_text" in event.type:
+            rich.print(event)
+
+    rich.print(stream.get_final_response())

examples/responses/background_streaming_async.py

@@ -0,0 +1,53 @@
+import asyncio
+from typing import List
+
+import rich
+from pydantic import BaseModel
+
+from openai import AsyncOpenAI
+
+
+class Step(BaseModel):
+    explanation: str
+    output: str
+
+
+class MathResponse(BaseModel):
+    steps: List[Step]
+    final_answer: str
+
+
+async def main() -> None:
+    client = AsyncOpenAI()
+    id = None
+    async with client.responses.stream(
+        input="solve 8x + 31 = 2",
+        model="gpt-4o-2024-08-06",
+        text_format=MathResponse,
+        background=True,
+    ) as stream:
+        async for event in stream:
+            if event.type == "response.created":
+                id = event.response.id
+            if "output_text" in event.type:
+                rich.print(event)
+            if event.sequence_number == 10:
+                break
+
+    print("Interrupted. Continuing...")
+
+    assert id is not None
+    async with client.responses.stream(
+        response_id=id,
+        starting_after=10,
+        text_format=MathResponse,
+    ) as stream:
+        async for event in stream:
+            if "output_text" in event.type:
+                rich.print(event)
+
+        rich.print(stream.get_final_response())
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "openai"
-version = "1.81.0"
+version = "1.82.0"
 description = "The official Python library for the openai API"
 dynamic = ["readme"]
 license = "Apache-2.0"