William Easton
William Easton
second iteration https://github.com/jlowin/fastmcp/pull/177
My guess is making it a fifo queue would be the best bet along with invoking a publish when the queue utilization hits a certain size like 50%
@legrego thoughts on fixing this in my current testing pr vs fixing later?
In addition, we can actually read the docstrings and use the values for tests. So we can get three things in one: 1. Documentation of what we expect from Elasticsearch...
I may have lied about being able to use the docstrings in the tests, using a better context manager is probably the better idea
``` class DatastreamService(BaseElasticsearchService): """Service for managing and querying Elasticsearch datastream lifecycle and metadata.""" async def get( self, name: str, filter_path: list[str] | None = None ) -> dict[Literal["data_streams"], list[dict[str, Any]]]:...
The nw server works just fine for me. I am able to get a forecast via the httpx client. Without the exact error you're seeing it will be hard to...
@andrewkroh and I will be chatting today but I think we should definitely chat about using the OTel input
Are you still able to reproduce this? I think this may have been fixed with https://github.com/modelcontextprotocol/python-sdk/pull/841 I cant reproduce this on main -- i see the exception on the server...
Just don't look too closely at how I'm enabling/disabling tools via settings -- that'll get fixed when the tags for enabled/disabled tools is implemented :D