Support custom certificate injection in all of watsonx.ai

Many large legacy IT organizations like ours don't use public certificates for internal services, but opt to use their own internal certificates, signed by their own Root CA.

Of course, by default, these certificates will be viewed as compromised by stock software such as watsonx.ai, as the Root CA certs are not in the standard trust store.

Watson Studio has a feature that supports injecting the root ca certificates, see Creating a secret to store shared custom certificates - IBM Documentation

This works fine and allows developers to use SSL verification as normally when communicating from python notebooks with internal services, such as internal databases and S3.

However, inexplicably, even though outside communication implies that Studio and Machine Learning are part of one unified product, share GUI elements, and in the minds of our developers, who are not familiar with previous re-branding permutations, are essentially just two facets of the same thing, WML does not support this feature as of 5.2.0.

So any reasonable person would expect that the same code if working in a notebook in Watson Studio, will also work when deployed as a Python function asset. The GUI strongly implies this, there is literally a button to promote assets.

Of course, this is not the case, as WS and WML are still developed by different teams that prioritize different general directions. WML simply doesn't inject the certificates and the SSL validation fails.

In this AHA we ask to add the certificate injection feature into WML. Generally, we ask to include tests that reflect the natural intutition that code that works in "projects" should also work in "deployment spaces".