fix: Streaming support detection for main LLM initialization

[Pouyanpi]

Problem

The main_llm_supports_streaming flag was not being set properly when the main LLM was initialized from config, causing streaming functionality to fail in the develop branch while it worked in v0.14.0.

Root Cause

During LLM initialization refactoring (#1221), the streaming support detection logic was only applied when iterating through config models, but was skipped when the main LLM was initialized directly from config in the new code path.

Fix

• Added streaming support detection when main LLM is initialized from config
• Added streaming support detection when LLM is provided via constructor
• Ensures main_llm_supports_streaming is properly set in both initialization paths

Test Plan

Added three tests to tests/test_streaming.py that verify main_llm_supports_streaming is correctly set:

1. test_main_llm_supports_streaming_flag_with_config(): LLM from config with streaming enabled
2. test_main_llm_supports_streaming_flag_with_constructor(): LLM via constructor with streaming enabled
3. test_main_llm_supports_streaming_flag_disabled_when_no_streaming(): Streaming disabled

These tests fail before this fix and pass after, preventing future regressions.

[Copilot]

Pull Request Overview

This pull request fixes an issue with the streaming support detection for the main LLM during initialization, ensuring that the streaming flag is properly set whether the LLM is created from configuration or via the constructor. Key changes include updating tests to verify streaming detection, adding a custom streaming provider fixture for testing, and refactoring streaming configuration logic in LLMRails.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
tests/test_streaming.py	Updated tests on streaming support with custom streaming providers and added tests for constructor-based LLM initialization.
nemoguardrails/rails/llm/llmrails.py	Introduced a dedicated _configure_main_llm_streaming function and updated its invocation in multiple initialization paths to properly set the streaming flag.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 91.66667% with 1 line in your changes missing coverage. Please review.

Project coverage is 69.57%. Comparing base (426d704) to head (728f249).

Files with missing lines	Patch %	Lines
nemoguardrails/rails/llm/llmrails.py	91.66%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1258      +/-   ##
===========================================
+ Coverage    69.54%   69.57%   +0.03%     
===========================================
  Files          161      161              
  Lines        16016    16023       +7     
===========================================
+ Hits         11138    11148      +10     
+ Misses        4878     4875       -3     

Flag	Coverage Δ
python	69.57% <91.66%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
nemoguardrails/rails/llm/llmrails.py	88.83% <91.66%> (+0.64%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[tgasser-nv]

Approved - please add a few more tests as requested for coverage

[Pouyanpi]

Thank you @tgasser-nv , I've added the tests and made the suggested changes.