Repository navigation
Conversation
joerowell
requested review from
aarnphm,
bbrowning,
chaunceyjiang and
sfeng33
as code owners
June 30, 2026 14:18
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
Contributor
|
This pull request has merge conflicts that must be resolved before it can be |
Re-implements the poolside_v1 tool parser (poolside Laguna) on the declarative parser engine instead of the standalone regex parser. Laguna emits the GLM-4.7 tool-call format verbatim (<tool_call>name<arg_key>k</arg_key><arg_value>v</arg_value> </tool_call>, function name inline with no newline separator), so it is backed by a new PoolsideParser whose poolside_v1_config mirrors glm47_moe_config. This inherits the engine's incremental string streaming and schema-aware argument coercion, and retires ~590 lines of bespoke parsing. - vllm/parser/poolside.py: PoolsideParser + poolside_v1_config (own config, structurally identical to glm47_moe_config). - vllm/parser/engine/registered_adapters.py: make_adapters(PoolsideParser). - vllm/tool_parsers/poolside_v1_tool_parser.py: thin PoolsideV1ToolParser (registry name unchanged). Scoped to the tool parser; reasoning stays on PoolsideV1ReasoningParser, which scopes the </think> search to the current assistant turn (Laguna-specific behaviour the generic engine reasoning parser does not provide). Equivalence with glm47 confirmed empirically: serving the Laguna-XS-2.1 checkpoint with --tool-call-parser glm47 parses tool calls cleanly and scores BFCL non-live simple_python 93.0% / multiple 95.5% (in line with bf16). Signed-off-by: Joe Rowell <joerowell4@gmail.com>
joerowell
force-pushed
the
joerowell/poolside-parser-engine
branch
from
July 9, 2026 14:48
864d084 to
9e619a5
Compare
Contributor
|
This pull request has merge conflicts that must be resolved before it can be |
|
This pull request has been automatically marked as stale because it has not had any activity within 90 days. It will be automatically closed if no further activity occurs within 30 days. Leave a comment if you feel this pull request should remain open. Thank you! |
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Re-implements the poolside_v1 tool parser (poolside Laguna) on the declarative parser engine instead of the standalone regex parser. Laguna emits the GLM-4.7 tool-call format verbatim (<tool_call>name<arg_key>k</arg_key><arg_value>v</arg_value> </tool_call>, function name inline with no newline separator), so it is backed by a new PoolsideParser whose poolside_v1_config mirrors glm47_moe_config. This inherits the engine's incremental string streaming and schema-aware argument coercion. This was initially suggested to us by @sfeng33 in #46486.
Purpose
Test Plan
Test Result
Essential Elements of an Effective PR Description Checklist
supported_models.mdandexamplesfor a new model.