anthropics / anthropics/anthropic-sdk-python
API "Streaming is required" error message is outdated
- 主要语言
- Python
- 星标
- 3.9k
- 派生
- 853
- 平均合并
- 1 天 11 小时
- 30 天内合并 PR
- 10
描述
Providing a max_tokens parameter greater than ~20k tokens without using the streaming API produces an error which references "https://github.com/anthropics/anthropic-sdk-python#long-requests", a heading which no longer exists since [0e409217](https://github.com/anthropics/anthropic-sdk-python/commit/0e4092173fc52d773675f8e3a4ced6a5e2f67c9f).
Two points:
1. This error message should be re-pointed to wherever the documentation currently lives
2. What is the right thing to do here, if for whatever reason we cannot use the streaming API? It no longer seems possible to omit the max_tokens field when making an API request, so I assume the prior advice is no longer appropriate. I can call _calculate_nonstreaming_timeout to determine the 'safe non-streaming maximum' but that's obviously brittle.
Full error:
```
File "/proj/api.py", line 615, in _execute_request
text, inp, out, cached = await self._execute_anthropic(request, model_id, self._get_anthropic())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/proj/api.py", line 441, in _execute_anthropic
response = await client.messages.create(**kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/.venv/lib64/python3.14/site-packages/anthropic/resources/messages/messages.py", line 2425, in create
timeout = self._client._calculate_nonstreaming_timeout(
max_tokens, MODEL_NONSTREAMING_TOKENS.get(model, None)
)
File "/.venv/lib64/python3.14/site-packages/anthropic/_base_client.py", line 732, in _calculate_nonstreaming_timeout
raise ValueError(
...<2 lines>...
)
ValueError: Streaming is required for operations that may take longer than 10 minutes. See https://github.com/anthropics/anthropic-sdk-python#long-requests for more details
```
This is with v0.88, but as far as I can tell there were no changes to the relevant documentation files since then.
贡献指南
评估
这个 Issue 还没有评估数据。