refactor(logger): deeply nested objects fail to show up in logs.

[Hweinstock]

Problem

Deeply nested objects fail to show up in the logger due to use of %O instead of JSON.stringify. Some customers use this information for debugging or other purposes.
Ex.

2024-11-12 13:23:39.682 [info] request from tab: tab-1 conversationID: 619df9a2-3ab2-4676-9896-9eb608d80582 request: {
  conversationState: {
    currentMessage: { userInputMessage: [Object] },
    chatTriggerType: 'MANUAL',
    customizationArn: undefined
  }
}

Solution

• Change codewhisper chat request/response log messages (which are very deeply nested) back to JSON.stringify.

---

License: I confirm that my contribution is made under the terms of the Apache 2.0 license.

[github-actions]

• This pull request modifies code in src/* but no tests were added/updated.
  • Confirm whether tests should be added or ensure the PR description explains why tests are not required.

[justinmk3]

should we skip the partialClone for the non-web case?

I think inspect() probably doesn't need to support partiaClone features (such as omitKeys) Because it's trivial for any caller to use inspect(partialClone(...)).

[Hweinstock]

I was worried that some of the options would then only apply in certain cases. Ex. if not on web and we pass omitKeys, it does nothing w/o error or warning. But I guess the caller can just wrap it themselves.

[justinmk3]

Ex. if not on web and we pass omitKeys,

inspect() doesn't really need to support omitKeys though. Unless that is a feature of nodejs inspect().

[justinmk3]

inspect doesn't need this options, because callers can easily use inspect(partialClone(o)) to opt-in if they want (and that's a good general practice: composability instead of "denormalization").

The use of partialClone for the "web" case in our inspect wrapper, is simply a way for us to support depth.

[justinmk3]

should this be mentioned in the lint rule?