minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[model-gateway] bug fix on module name (#16332)

Browse Source
This commit is contained in:
Chang Su
2026-01-02 21:56:18 -08:00
committed by GitHub
parent bdde949619
commit d8ac5eecf7
66 changed files with 189 additions and 222 deletions
Download Patch File Download Diff File Expand all files Collapse all files
sgl-model-gateway/bindings/golang/src/client.rs
+5 -5
View File
@@ -8,11 +8,11 @@ use tokio::runtime::Runtime;
use once_cell::sync::Lazy;
use uuid::Uuid;
use sgl_model_gateway::tokenizer::create_tokenizer_from_file;
use sgl_model_gateway::tokenizer::traits::Tokenizer;
use sgl_model_gateway::grpc_client::sglang_scheduler::SglangSchedulerClient;
use sgl_model_gateway::protocols::chat::ChatCompletionRequest;
use sgl_model_gateway::routers::grpc::utils::{process_chat_messages, generate_tool_constraints};
use smg::tokenizer::create_tokenizer_from_file;
use smg::tokenizer::traits::Tokenizer;
use smg::grpc_client::sglang_scheduler::SglangSchedulerClient;
use smg::protocols::chat::ChatCompletionRequest;
use smg::routers::grpc::utils::{process_chat_messages, generate_tool_constraints};
use super::error::{SglErrorCode, set_error_message};
use super::grpc_converter::sgl_grpc_response_converter_create;
sgl-model-gateway/bindings/golang/src/grpc_converter.rs
+17 -17
View File
@@ -9,12 +9,12 @@ use serde_json::Value;
use tokio::runtime::Runtime;
use once_cell::sync::Lazy;
use sgl_model_gateway::tokenizer::traits::Tokenizer;
use sgl_model_gateway::tokenizer::stream::DecodeStream;
use sgl_model_gateway::tool_parser::ToolParser;
use sgl_model_gateway::protocols::common::{Tool, ToolChoice, ToolChoiceValue, ToolCallDelta, FunctionCallDelta, Usage, StringOrArray};
use sgl_model_gateway::tokenizer::stop::StopSequenceDecoder;
use sgl_model_gateway::grpc_client::sglang_proto as proto;
use smg::tokenizer::traits::Tokenizer;
use smg::tokenizer::stream::DecodeStream;
use smg::tool_parser::ToolParser;
use smg::protocols::common::{Tool, ToolChoice, ToolChoiceValue, ToolCallDelta, FunctionCallDelta, Usage, StringOrArray};
use smg::tokenizer::stop::StopSequenceDecoder;
use smg::grpc_client::sglang_proto as proto;
use super::error::{SglErrorCode, set_error_message, clear_error_message};
use super::tokenizer::TokenizerHandle;
@@ -22,9 +22,9 @@ use super::utils::generate_tool_call_id;
/// Global parser factory (initialized once)
// Use the re-exported ParserFactory from tool_parser module
static PARSER_FACTORY: Lazy<sgl_model_gateway::tool_parser::ParserFactory> = Lazy::new(|| {
static PARSER_FACTORY: Lazy<smg::tool_parser::ParserFactory> = Lazy::new(|| {
// ParserFactory is re-exported from tool_parser::factory, so we can use it directly
sgl_model_gateway::tool_parser::ParserFactory::default()
smg::tool_parser::ParserFactory::default()
});
/// Global tokio runtime for async operations
@@ -151,7 +151,7 @@ pub unsafe extern "C" fn sgl_grpc_response_converter_create(
// Create stop decoder if needed
let stop_decoder = if stop.is_some() || stop_token_ids.is_some() {
Some(Arc::new(tokio::sync::Mutex::new(
sgl_model_gateway::routers::grpc::utils::create_stop_decoder(
smg::routers::grpc::utils::create_stop_decoder(
&tokenizer,
stop.as_ref(),
stop_token_ids.as_ref(),
@@ -389,9 +389,9 @@ pub(crate) async fn convert_proto_chunk_to_openai(
request_id: &str,
created: u64,
system_fingerprint: Option<&str>,
) -> Result<Option<sgl_model_gateway::protocols::chat::ChatCompletionStreamResponse>, String> {
use sgl_model_gateway::grpc_client::sglang_proto::generate_response::Response::*;
use sgl_model_gateway::protocols::chat::{ChatCompletionStreamResponse, ChatMessageDelta, ChatStreamChoice};
) -> Result<Option<smg::protocols::chat::ChatCompletionStreamResponse>, String> {
use smg::grpc_client::sglang_proto::generate_response::Response::*;
use smg::protocols::chat::{ChatCompletionStreamResponse, ChatMessageDelta, ChatStreamChoice};
match proto_response.response {
Some(Chunk(chunk)) => {
@@ -427,19 +427,19 @@ pub(crate) async fn convert_proto_chunk_to_openai(
let mut text = String::new();
for &token_id in &chunk.token_ids {
match decoder_guard.process_token(token_id).unwrap_or_else(|_| {
sgl_model_gateway::tokenizer::stop::SequenceDecoderOutput::Held
smg::tokenizer::stop::SequenceDecoderOutput::Held
}) {
sgl_model_gateway::tokenizer::stop::SequenceDecoderOutput::Text(t) => {
smg::tokenizer::stop::SequenceDecoderOutput::Text(t) => {
text.push_str(&t);
}
sgl_model_gateway::tokenizer::stop::SequenceDecoderOutput::StoppedWithText(t) => {
smg::tokenizer::stop::SequenceDecoderOutput::StoppedWithText(t) => {
text.push_str(&t);
break;
}
sgl_model_gateway::tokenizer::stop::SequenceDecoderOutput::Stopped => {
smg::tokenizer::stop::SequenceDecoderOutput::Stopped => {
break;
}
sgl_model_gateway::tokenizer::stop::SequenceDecoderOutput::Held => {}
smg::tokenizer::stop::SequenceDecoderOutput::Held => {}
}
}
text
sgl-model-gateway/bindings/golang/src/postprocessor.rs
+1 -1
View File
@@ -14,7 +14,7 @@ use std::ptr;
use std::sync::Arc;
use serde_json::Value;
use sgl_model_gateway::grpc_client::sglang_proto as proto;
use smg::grpc_client::sglang_proto as proto;
use super::error::{SglErrorCode, set_error_message};
use super::grpc_converter::GrpcResponseConverterHandle;
sgl-model-gateway/bindings/golang/src/preprocessor.rs
+3 -3
View File
@@ -12,9 +12,9 @@ use std::os::raw::{c_char, c_int};
use std::ptr;
use std::os::raw::c_uint;
use sgl_model_gateway::tokenizer::create_tokenizer_from_file;
use sgl_model_gateway::protocols::chat::ChatCompletionRequest;
use sgl_model_gateway::routers::grpc::utils::{process_chat_messages, generate_tool_constraints};
use smg::tokenizer::create_tokenizer_from_file;
use smg::protocols::chat::ChatCompletionRequest;
use smg::routers::grpc::utils::{process_chat_messages, generate_tool_constraints};
use super::error::{SglErrorCode, set_error_message};
use super::memory::{sgl_free_string, sgl_free_token_ids};
sgl-model-gateway/bindings/golang/src/stream.rs
+1 -1
View File
@@ -23,7 +23,7 @@ use tokio::runtime::Runtime;
use once_cell::sync::Lazy;
use futures_util::StreamExt;
use sgl_model_gateway::grpc_client::{sglang_proto as proto, sglang_scheduler::{SglangSchedulerClient, AbortOnDropStream}};
use smg::grpc_client::{sglang_proto as proto, sglang_scheduler::{SglangSchedulerClient, AbortOnDropStream}};
use super::error::{SglErrorCode, set_error_message};
use super::grpc_converter::{GrpcResponseConverterHandle, convert_proto_chunk_to_openai};
sgl-model-gateway/bindings/golang/src/tokenizer.rs
+1 -1
View File
@@ -6,7 +6,7 @@ use std::ptr;
use std::sync::Arc;
use serde_json::Value;
use sgl_model_gateway::tokenizer::{
use smg::tokenizer::{
create_tokenizer_from_file,
traits::Tokenizer as TokenizerTrait,
chat_template::ChatTemplateParams,
sgl-model-gateway/bindings/golang/src/tool_parser.rs
+2 -2
View File
@@ -9,8 +9,8 @@ use serde_json::{json, Value};
use tokio::runtime::Runtime;
use once_cell::sync::Lazy;
use sgl_model_gateway::tool_parser::{ParserFactory, ToolParser};
use sgl_model_gateway::protocols::common::Tool;
use smg::tool_parser::{ParserFactory, ToolParser};
use smg::protocols::common::Tool;
use super::error::{SglErrorCode, set_error_message, clear_error_message};
use super::utils::generate_tool_call_id;
sgl-model-gateway/bindings/python/src/lib.rs
+1 -1
View File
@@ -1,5 +1,5 @@
use pyo3::prelude::*;
use sgl_model_gateway::*;
use smg::*;
use once_cell::sync::OnceCell;
use std::collections::HashMap;