Add Support for Simulated Streaming Toggle and Provider Compatibility Checks

[tuannvm]

Summary:

This PR introduces configuration options and logic to handle simulated streaming, allowing flexibility when native streaming support is unavailable from certain providers. It adds an environment-aware support check for multiple providers to prevent misconfiguration, ensuring the application only proceeds with simulated streaming when appropriate.

Provider	Native Streaming Support	Behavior with --enable-simulated-streaming=true (default)	Behavior with --enable-simulated-streaming=false
openai	No	Works with simulated streaming	Fails at startup with error message
grok	No	Works with simulated streaming	Fails at startup with error message
gemini	Yes	Works with native streaming	Works with native streaming
ollama	No	Works with native streaming	Fails at startup with error message
llamacpp	No	Works with native streaming	Fails at startup with error message
azopenai	No	Works with native streaming	Fails at startup with error message
unknown	Assumed No	Works with simulated streaming	Fails at startup with error message

When the application starts:

1. If --enable-simulated-streaming=false, the code checks if the selected provider supports native streaming
2. If native streaming is not supported, it fails immediately with an error message
3. If native streaming is supported, it sets environment variables (KUBECTL_AI_REQUIRE_NATIVE_STREAMING=true and provider-specific variables) and continues

This approach ensures that applications that depend on real-time token-by-token streaming for UI updates or early termination can enforce this requirement at startup time.

Affected Modules:

• cmd/main.go (Options definition, CLI flag binding, environment setup)
• main logic flow (runtime streaming capability validation)
• provider support detection (checkNativeStreamingSupport function)

Key Details:

• Adds a new CLI flag --enable-simulated-streaming with a default value of true.
• Implements checkNativeStreamingSupport() to determine if a provider natively supports streaming features, with default assumptions.
• During runtime, if --enable-simulated-streaming=false is specified, the code verifies provider support before proceeding.
• Ensures users are warned about provider limitations and configuration mismatches, preventing runtime errors.

Potential Impacts:

• Users can explicitly disable simulated streaming, but only if their provider supports native streaming.
• Prevents misconfiguration by aborting excessive unsupported setup attempts.
• Enhances robustness across multiple providers with varying streaming capabilities.
• Adds flexibility for future providers by expanding support checks in checkNativeStreamingSupport().

Testing Steps:

• Run with --enable-simulated-streaming=false against providers known to support native streaming (gemini), verify normal operation.
• Run with --enable-simulated-streaming=false against providers without native streaming support (azopenai, grok, llamacpp, ollama, openai) and confirm proper error handling.
• Verify default behavior with --enable-simulated-streaming omitted, ensuring it defaults to allowing simulated streaming.
• Confirm CLI flag parsing correctly sets the configuration and influences runtime validation.
• Test warning logs for unknown providers to ensure future compatibility.

[zvdy]

Just realized this was open, perhaps #215 solves what is mentioned here

[droot]

Just realized this was open, perhaps #215 solves what is mentioned here

I think so too. Awesome to see the progress on the streaming (and auto-detection/fallback).