Skip to content

fix(overflow): add hub GLM phrase to context-overflow detector #314

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
linj-glitch merged 1 commit into from
Aug 5, 2026
+14 −0
Merged

fix(overflow): add hub GLM phrase to context-overflow detector #314

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions crates/libsy-llm-client/src/backend.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,6 +24,7 @@ const OPENAI_OVERFLOW_PHRASES: &[&str] = &[
"context window",
"context length is only",
"please reduce the length of the input",
"exceeds the maximum allowed input length",
];

// Anthropic has no structured `error.code`, so detection is phrase-based only.
Expand Down Expand Up @@ -310,6 +311,10 @@ mod tests {
r#"{"error":{"message":"the model's context length is only 131072 tokens"}}"#
));
assert!(!backend.is_context_overflow(r#"{"error":{"code":"invalid_api_key"}}"#));
// Hub GLM (LiteLLM-wrapped): code is "400", detection relies on phrase match.
assert!(backend.is_context_overflow(
r#"{"error":{"message":"Input length 877338 exceeds the maximum allowed input length of 639968 tokens","code":"400"}}"#
));
}

#[test]
Expand Down
9 changes: 9 additions & 0 deletions crates/switchyard-components/src/backends/openai.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -528,6 +528,7 @@ const OPENAI_OVERFLOW_PHRASES: &[&str] = &[
"context window",
"context length is only",
"please reduce the length of the input",
"exceeds the maximum allowed input length",
];

fn is_context_overflow(body: &str) -> bool {
Expand Down Expand Up @@ -759,6 +760,14 @@ mod tests {
assert!(!is_context_overflow(body));
}

#[test]
fn context_overflow_hub_glm_matches() {
// Hub GLM error via LiteLLM: code is "400" (not context_length_exceeded),
// so detection relies on phrase matching.
let body = r#"{"error":{"message":"Input length 877338 exceeds the maximum allowed input length of 639968 tokens","code":"400"}}"#;
assert!(is_context_overflow(body));
}

#[test]
fn context_overflow_nvidia_litellm_wrap_matches() {
// Body shape observed from inference-api.nvidia.com's LiteLLM proxy
Expand Down
Loading