Run Langwatch locally can't find optimization results in localhost endpoint

[hanhanwu]

When I was running this code using endpoint https://app.langwatch.ai, I could see prompt optimization results got generated automatically:

Same code, I just changed endpoint to localhost (http://localhost:5560), the code can run successfully, all shown here, but the dashboard of my localhost can't show the chart I got above. It's empty Langwatch dashboard...

Do you know how can I show the same chart got above in http://localhost:5560?

[rogeriochaves]

hey @hanhanwu!

sorry for the delay in replying, but we changed quite a bunch of things recently, both on the local docker set up and our sdk. I see though there there is something a bit odd on your jupyter notebook:

langwatch.dspy.__init__(experiment="hanhan_exp1", optimizer=optimizer)

here it should be langwatch.dspy.init instead of __init__, it's actually a bit surprising if it worked on the cloud version

can you try:

• make sure you are using our sdk on langwatch==0.2.2 or higher
• pull in again the latest main/docker image when running locally
• replace langwatch.dspy.__init__ with ``langwatch.dspy.init`

and let me know if now it works for you?

[hanhanwu]

Hi @rogeriochaves

I did update to the latest Langwatch, but same error, if using langwatch.dspy.init() got error below; if using langwatch.dspy.__init__(), after running compile succeeded, there's nothing shown in http://localhost:5560/ dashboard

I'm using Windows 11. Are you guys able to test on Windows machine and see whether the local version can update prompt opt results?