[router] fix chat template loading and tokenizer path (#10999)

This commit is contained in:
Simo Lin
2025-09-27 23:54:12 -04:00
committed by GitHub
parent 72392f2908
commit 5519766a4d
5 changed files with 418 additions and 9 deletions

View File

@@ -134,6 +134,36 @@ fn is_likely_sentencepiece(buffer: &[u8]) -> bool {
|| buffer.windows(4).any(|w| w == b"</s>"))
}
/// Helper function to discover chat template files in a directory
pub fn discover_chat_template_in_dir(dir: &Path) -> Option<String> {
use std::fs;
// Priority 1: Look for chat_template.json (contains Jinja in JSON format)
let json_template_path = dir.join("chat_template.json");
if json_template_path.exists() {
return json_template_path.to_str().map(|s| s.to_string());
}
// Priority 2: Look for chat_template.jinja (standard Jinja file)
let jinja_path = dir.join("chat_template.jinja");
if jinja_path.exists() {
return jinja_path.to_str().map(|s| s.to_string());
}
// Priority 3: Look for any .jinja file (for models with non-standard naming)
if let Ok(entries) = fs::read_dir(dir) {
for entry in entries.flatten() {
if let Some(name) = entry.file_name().to_str() {
if name.ends_with(".jinja") && name != "chat_template.jinja" {
return entry.path().to_str().map(|s| s.to_string());
}
}
}
}
None
}
/// Factory function to create tokenizer from a model name or path (async version)
pub async fn create_tokenizer_async(
model_name_or_path: &str,
@@ -161,14 +191,32 @@ pub async fn create_tokenizer_async(
// Look for tokenizer.json in the cache directory
let tokenizer_path = cache_dir.join("tokenizer.json");
if tokenizer_path.exists() {
create_tokenizer_from_file(tokenizer_path.to_str().unwrap())
// Try to find a chat template file in the cache directory
let chat_template_path = discover_chat_template_in_dir(&cache_dir);
let tokenizer_path_str = tokenizer_path.to_str().ok_or_else(|| {
Error::msg(format!(
"Tokenizer path is not valid UTF-8: {:?}",
tokenizer_path
))
})?;
create_tokenizer_with_chat_template(
tokenizer_path_str,
chat_template_path.as_deref(),
)
} else {
// Try other common tokenizer file names
let possible_files = ["tokenizer_config.json", "vocab.json"];
for file_name in &possible_files {
let file_path = cache_dir.join(file_name);
if file_path.exists() {
return create_tokenizer_from_file(file_path.to_str().unwrap());
let chat_template_path = discover_chat_template_in_dir(&cache_dir);
let file_path_str = file_path.to_str().ok_or_else(|| {
Error::msg(format!("File path is not valid UTF-8: {:?}", file_path))
})?;
return create_tokenizer_with_chat_template(
file_path_str,
chat_template_path.as_deref(),
);
}
}
Err(Error::msg(format!(