[FRONTEND] OpenAI tools support named functions

[br3no]

PARTIAL FIX #1869 (and #5008) -- only named tools

BEFORE SUBMITTING, PLEASE READ THE CHECKLIST BELOW AND FILL IN THE DESCRIPTION ABOVE

---

PR Checklist (Click to Expand)

Thank you for your contribution to vLLM! Before submitting the pull request, please ensure the PR meets the following criteria. This helps vLLM maintain the code quality and improve the efficiency of the review process.

PR Title and Classification

Only specific types of PRs will be reviewed. The PR title is prefixed appropriately to indicate the type of change. Please use one of the following:

• [Bugfix] for bug fixes.
• [CI/Build] for build or continuous integration improvements.
• [Doc] for documentation fixes and improvements.
• [Model] for adding a new model or improving an existing model. Model name should appear in the title.
• [Frontend] For changes on the vLLM frontend (e.g., OpenAI API server, LLM class, etc.)
• [Kernel] for changes affecting CUDA kernels or other compute kernels.
• [Core] for changes in the core vLLM logic (e.g., LLMEngine, AsyncLLMEngine, Scheduler, etc.)
• [Hardware][Vendor] for hardware-specific changes. Vendor name should appear in the prefix (e.g., [Hardware][AMD]).
• [Misc] for PRs that do not fit the above categories. Please use this sparingly.

Note: If the PR spans more than one category, please include all relevant prefixes.

Code Quality

The PR need to meet the following code quality standards:

• We adhere to Google Python style guide and Google C++ style guide.
• Pass all linter checks. Please use format.sh to format your code.
• The code need to be well-documented to ensure future contributors can easily understand the code.
• Include sufficient tests to ensure the project to stay correct and robust. This includes both unit tests and integration tests.
• Please add documentation to docs/source/ if the PR modifies the user-facing behaviors of vLLM. It helps vLLM user understand and utilize the new features or changes.

Notes for Large Changes

Please keep the changes as concise as possible. For major architectural changes (>500 LOC excluding kernel/data/config/test), we would expect a GitHub issue (RFC) discussing the technical design and justification. Otherwise, we will tag it with rfc-required and might not go through the PR.

What to Expect for the Reviews

The goal of the vLLM team is to be a transparent reviewing machine. We would like to make the review process transparent and efficient and make sure no contributor feel confused or frustrated. However, the vLLM team is small, so we need to prioritize some PRs over others. Here is what you can expect from the review process:

• After the PR is submitted, the PR will be assigned to a reviewer. Every reviewer will pick up the PRs based on their expertise and availability.
• After the PR is assigned, the reviewer will provide status update every 2-3 days. If the PR is not reviewed within 7 days, please feel free to ping the reviewer or the vLLM team.
• After the review, the reviewer will put an action-required label on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.
• Please respond to all comments within a reasonable time frame. If a comment isn't clear or you disagree with a suggestion, feel free to ask for clarification or discuss the suggestion.

Thank You

Finally, thank you for taking the time to read these guidelines and for your interest in contributing to vLLM. Your contributions make vLLM a great tool for everyone!

[br3no]

@simon-mo

This PR implements only the "named tool" feature of the OpenAI API, as sketched in my comment here: #1869 (comment)

This exposes the guided_json functionality in an OpenAI compatible way.

Unfortunately, I have branched this change out of another small open PR (#5029), regarding correct logprob formats. Let me know if this is too confusing and I'll separate the two again.

[simon-mo]

Thank you for this PR. Will review and merge once the logprobs is in. Let's get this in.

[K-Mistele]

#4656 is a more complete implementation of this that supports auto tool choice (for models that support it); just need to fix jinja templating for the tool usage system prompt since different models have different system prompt formats (e.g. mistral 7B instruct v0.3, vs. Nous Hermes' Hermes 2 Pro Llama 3 8B)

[br3no]

@K-Mistele, I'm aware of #4656. I believe that a step-by-step approach will more easily find it's way into vLLM, so starting with the smallest possible increment is a much easier and more manageable task.

Once this is merged, #4656 can continue with the remaining missing features.

[br3no]

@simon-mo this is ready for review now.

[br3no]

@simon-mo the failing tests seem to be infrastructure-related and unrelated with the change. I hope you have time to review this soon. I’d love to have this in the next release. Let me know if you have any feedback so I can address all issues before you cut the next tag!

[simon-mo]

This is great and thank you for breaking up the big PR. Most comments are cosmetic. Just to confirm, we are not appending the schema to the prompt and expecting users to do it? Should we add it in?

[br3no]

@simon-mo cosmetic changes pushed.

Yes, correct, we are not appending the schema to the prompt.

This question will become more relevant once we add support for auto and required. Since this is highly dependent on the model and the use-case (e.g. language of the prompt), we will need to add support for configurable tool-prompt templates (see #4656, where @K-Mistele proposed just that, which I think is a good idea).

I think it's fair to have the user take care of this in this first iteration.

What do you think? I could add a documentation on the topic to make sure people know how to use this.

[br3no]

I added a suggestion for the documentation. Let me know what you think.

[br3no]

Outlines or lm-format-enforcer guarantee that for us.