chore: [testing] Mic/todos in moderators fix

[michaelneale]

adding in fixes from #54 to try to fix broken test on #52

Also, we can't default to contextsummarizer while it hard codes gpt4o #57

[codefromthecrypt]

🚢

[codefromthecrypt]

actually this has the wrong model thing again..

>               raise httpx.HTTPStatusError(f"{e}\n{response.text}", request=e.request, response=e.response)
E               httpx.HTTPStatusError: Client error '404 Not Found' for url 'http://localhost:11434/v1/chat/completions'
E               For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/404
E               {"error":{"message":"model \"gpt-4o-mini\" not found, try pulling it first","type":"api_error","param":null,"code":null}}

[michaelneale]

@codefromthecrypt hrm - yes same as last time. Is this same bug? I am totally confused

[codefromthecrypt]

let's remove this default so there's no chance of setting it wrong (e.g. better to error due to not inheriting model than using wrong one)

class ContextSummarizer(ContextTruncate):
    def __init__(
        self,
        model: str = "gpt-4o-mini",

[michaelneale]

yes, or it needs to be fixed to take in the correct model