fix: use a cache file name with a fixed size.

[matthisholleville]

Closes #348

📑 Description

Our current caching method creates a file with the following pattern:

$BACKEND_AI-$LANG-$SENC

SENC corresponds to a base64 encoding of the errors detected on an object during analysis.

This approach fails when we have multiple errors reported for the same object. The base64 string becomes too long, resulting in the following error:

error storing value to cache: open /Users/matthisholleville/Library/Caches/k8sgpt/openai-english-SW5ncmVzcyBrOHNncHQvZmFrZS1pbmdyZXNzIGRvZXMgbm90IHNwZWNpZnkgYW4gSW5ncmVzcyBjbGFzcy4gSW5ncmVzcyB1c2VzIHRoZSBzZXJ2aWNlIGs4c2dwdC9mYWtlLXNlcnZpY2Ugd2hpY2ggZG9lcyBub3QgZXhpc3QuIEluZ3Jlc3MgdXNlcyB0aGUgc2VjcmV0IGs4c2dwdC9mYWtlLXNlY3JldCBhcyBhIFRMUyBjZXJ0aWZpY2F0ZSB3aGljaCBkb2VzIG5vdCBleGlzdC4=: file name too long

To address this issue, we use SHA256 to encode filename and keep fixe file name size

✅ Checks

• My pull request adheres to the code style of this project
• My code requires changes to the documentation
• I have updated the documentation as required
• All the tests have passed

ℹ Additional Information

I have also fixed another bug related to the no-cache feature. Previously, when we activated no-cache, the cache was not used, but the command result was not updated either. This resulted in bugs like the following:

$ k8sgpt analyze --explain

0 k8sgpt/fake-hpa(fake-hpa)
Error: HorizontalPodAutoscaler uses StatefulSet/fake-deployment as ScaleTargetRef which does not exist.
Fake

$ k8sgpt analyze --explain --no-cache

0 k8sgpt/fake-hpa(fake-hpa)
Error: HorizontalPodAutoscaler uses StatefulSet/fake-deployment as ScaleTargetRef which does not exist.
The HorizontalPodAutoscaler is trying to scale a StatefulSet/fake-deployment, which does not exist.

$ k8sgpt analyze --explain

0 k8sgpt/fake-hpa(fake-hpa)
Error: HorizontalPodAutoscaler uses StatefulSet/fake-deployment as ScaleTargetRef which does not exist.
Fake

I have fixed this issue so that when the --no-cache option is enabled, the cache is not used, but it is updated with the latest response.

[thschue]

For me, this looks good ...

@AlexsJones : What do you think?

[thschue]

lgtm