fix: only enforce token limit in direct LLM mode

[Ahmath-Gadji]

Move the token-limit check so it only runs in direct LLM mode. In RAG mode the pipeline trims chat history to chat_history_depth (default 4) before calling the LLM, so the pre-trim check was falsely rejecting long conversations that would fit once truncated. Direct LLM mode still enforces the limit since no trimming happens there.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 3b3ba131-0ddc-4e14-8077-23dbbdea07ca

📥 Commits

Reviewing files that changed from the base of the PR and between a3f5070 and 2c04fd4.

📒 Files selected for processing (1)

• openrag/routers/openai.py

---

📝 Walkthrough

Walkthrough

The token limit validation in the OpenAI router has been relocated to execute conditionally only for direct LLM model requests, moving it from a universal pre-flight check to logic nested within the direct-model handling branch, skipping validation for RAG-partitioned models.

Changes

Cohort / File(s)	Summary
Token Limit Validation Restructure openrag/routers/openai.py	Moved check_tokens_limit() call from unconditional pre-flight execution to conditional execution within the direct LLM model branch, causing RAG-partitioned models to bypass token limit validation.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Poem

🐰 A token check hops to new ground,
Now direct models keep it around,
While RAG-partitioned paths skip ahead,
Control flow branches where logic is led! ✨

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: the token limit check is now only enforced in direct LLM mode, which matches the code modification that moved the check to run conditionally inside the direct-model branch.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/token_check

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}