[Fix] Unified LLM Whisperer adapters

[jagadeeswaran-zipstack]

What

• Remove the LLMWhisperer v1 adapter and add it as an option in the V2 adapter. The versions v1 and v2 should be a dropdown with v2 being the default.
  ...

Why

• For better UI/UX experience
  ...

How

...

Relevant Docs

Related Issues or PRs

Dependencies Versions / Env Variables

Notes on Testing

...

Screenshots

...

Checklist

I have read and understood the Contribution Guidelines.

[chandrasekharan-zipstack]

@jagadeeswaran-zipstack

1. please consult with @gaya3-zipstack and update the version in __init__.py
2. Please help address the ask of separating envs
3. What about removing the other v2 sub-package?

[chandrasekharan-zipstack]

LGTM - left a NIT comment around documentation

[muhammad-ali-e]

please incorporate the latest changes in llm_whisperer (v1 and v2) adapters into this PR

[muhammad-ali-e]

@jagadeeswaran-zipstack
NIT: To ensure a clean codebase with clear separation between versions—and to make it easy to deprecate and remove older versions without affecting newer ones—we can follow this approach:

1. Define a Common Base Class
   Introduce a BaseLLMWhisperer class to encapsulate shared functionality (e.g., retrieving a JSON schema). If no common logic is needed, we can skip this base class to keep versions fully independent.

2. Implement Versioned Classes (LLMWhispererV1 and LLMWhispererV2)
   Each version will inherit from X2TextAdapter and override get_id() and get_name(), ensuring minimal dependencies between versions. This separation allows V1 to be removed easily without impacting V2.

3. Create a Proxy Class (LLMWhispererProxy)
   This factory will handle version selection dynamically, ensuring that the appropriate implementation is used based on the requested version, further reinforcing modularity.

class LLMWhispererProxy:
    _versions = {
        "v1": LLMWhispererV1,
        "v2": LLMWhispererV2
    }

    def __init__(self, settings: Dict[str, Any]):
        self.settings = settings
        self.version = settings.get("version", "v1") 
        if self.version not in self._versions:
            raise ValueError(f"Unsupported LLMWhisperer version: {self.version}")
        self.adapter = self._versions[self.version](settings)

    def get_id(self) -> str:
        return self.adapter.get_id()

    def get_name(self) -> str:
        return self.adapter.get_name()
        .......
      <other methods>
        

metadata = {
    "name": "LLMWhisperer",
    "version": "...",
    "adapter": LLMWhispererProxy,
    "description": "LLMWhisperer X2Text adapter supporting multiple versions",
    "is_active": True,
}

[jagadeeswaran-zipstack]

@jagadeeswaran-zipstack NIT: To ensure a clean codebase with clear separation between versions—and to make it easy to deprecate and remove older versions without affecting newer ones—we can follow this approach:

1. Define a Common Base Class
   Introduce a BaseLLMWhisperer class to encapsulate shared functionality (e.g., retrieving a JSON schema). If no common logic is needed, we can skip this base class to keep versions fully independent.
2. Implement Versioned Classes (LLMWhispererV1 and LLMWhispererV2)
   Each version will inherit from X2TextAdapter and override get_id() and get_name(), ensuring minimal dependencies between versions. This separation allows V1 to be removed easily without impacting V2.
3. Create a Proxy Class (LLMWhispererProxy)
   This factory will handle version selection dynamically, ensuring that the appropriate implementation is used based on the requested version, further reinforcing modularity.

class LLMWhispererProxy:
    _versions = {
        "v1": LLMWhispererV1,
        "v2": LLMWhispererV2
    }

    def __init__(self, settings: Dict[str, Any]):
        self.settings = settings
        self.version = settings.get("version", "v1") 
        if self.version not in self._versions:
            raise ValueError(f"Unsupported LLMWhisperer version: {self.version}")
        self.adapter = self._versions[self.version](settings)

    def get_id(self) -> str:
        return self.adapter.get_id()

    def get_name(self) -> str:
        return self.adapter.get_name()
        .......
      <other methods>
        

metadata = {
    "name": "LLMWhisperer",
    "version": "...",
    "adapter": LLMWhispererProxy,
    "description": "LLMWhisperer X2Text adapter supporting multiple versions",
    "is_active": True,
}

Will make these improvements Iteratively.

[muhammad-ali-e]

@jagadeeswaran-zipstack , please ensure we provide a timeout or wait_timeout for the whisperer-client. This might be handled through the JSON schema, but I think we should set some upper and lower (default) caps on our side. We can implement this later if needed.
cc: @johnyrahul @hari-kuriakose

[jagadeeswaran-zipstack]

@jagadeeswaran-zipstack , please ensure we provide a timeout or wait_timeout for the whisperer-client. This might be handled through the JSON schema, but I think we should set some upper and lower (default) caps on our side. We can implement this later if needed. cc: @johnyrahul @hari-kuriakose

@muhammad-ali-e We do have wait timeout which has default value of 300s. refer

[johnyrahul]

Do we allow the user to set this from screen?

[johnyrahul]

or can we at least take this from env