Feature Request: Allow safety_settings to be passed into Vertex Gemini.

[LuisV1296]

I'm a software developer currently working on a implementation of llamaindexTS to work with several LLM implementations in a configurable platform. Recently I started making the implementation for VertexAI through your Gemini implementation.
The implementation is very straightforward until I received this error message trying to execute the .chat() method:
{"level":"error","message":"LLM error: [VertexAI.ClientError]: got status: 400 Bad Request. {"error":{"code":400,"message":"User has requested a restricted HarmBlockThreshold setting BLOCK_NONE. You can get access either (
a) through an allowlist via your Google account team, or (b) by switching your account type to monthly invoiced billing via this instruction: https://cloud.google.com/billing/docs/how-to/invoiced-billing.\",\"status\":\"INVALID_ARGUMENT\"}}","reqId":"67645e3e9ee275d83956c68e"}

After some research I realized that the BLOCK_NONE setting is now under a huge paywall so I started looking at the source code to realize that the Safety Settings are not configurable by any possible ways on your implementation, and which is worse, there's a DEFAULT_SAFETY_SETTINGS constant that is imported at line 37 in https://github.com/run-llama/LlamaIndexTS/blob/main/packages/llamaindex/src/llm/gemini/base.ts which is declared at https://github.com/run-llama/LlamaIndexTS/blob/main/packages/llamaindex/src/llm/gemini/utils.ts in line 319 that has all the HarmBlockThresholds levels set in BLOCK_NONE. This DEFAULT_SAFETY_SETTINGS is hardcoded in the implementation of the streamChat and nonStreamChat methods for Gemini (see lines 243, 247, 286 and 290 in https://github.com/run-llama/LlamaIndexTS/blob/main/packages/llamaindex/src/llm/gemini/base.ts) which makes impossible to configure those values unless making a custom implementation of those classes which drives to several other issues.

There's a related post for the python version of llama which is already been solved: run-llama/llama_index#10788.

My request is to simply allow the user to configure their own safety_settings with something like this:

const safety_settings: SafetySetting[] = [
            {
                category: HarmCategory.HARM_CATEGORY_HATE_SPEECH,
                threshold: HarmBlockThreshold.BLOCK_ONLY_HIGH,
            },
            {
                category: HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT,
                threshold: HarmBlockThreshold.BLOCK_ONLY_HIGH,
            },
            {
                category: HarmCategory.HARM_CATEGORY_HARASSMENT,
                threshold: HarmBlockThreshold.BLOCK_ONLY_HIGH,
            },
            {
                category: HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT,
                threshold: HarmBlockThreshold.BLOCK_ONLY_HIGH,
            },
        ];

and send them as a parameter on the Gemini Constructor class or using a function.

Thanks for your time.