Merge branch 'pr/happynewplay/41'

src/exec/fun_asr_nano.rs

@@ -27,7 +27,9 @@ impl ExecModel for FunASRNanoExec {
         println!("Time elapsed in load model is: {:?}", i_duration);
 
         // Create ChatCompletionParameters for ASR
-        let url = &input[1];
+        let url = input.get(1).ok_or_else(|| {
+            anyhow::anyhow!("fun-asr-nano requires a second input: audio path or URL")
+        })?;
         let input_url = if url.starts_with("http://")
             || url.starts_with("https://")
             || url.starts_with("file://")

src/exec/glm_asr_nano.rs

@@ -28,7 +28,9 @@ impl ExecModel for GlmASRNanoExec {
 
         // Create ChatCompletionParameters for ASR
         // Input should be an audio file path
-        let url = &input[1];
+        let url = input.get(1).ok_or_else(|| {
+            anyhow::anyhow!("glm-asr-nano requires a second input: audio path or URL")
+        })?;
         let input_url = if url.starts_with("http://")
             || url.starts_with("https://")
             || url.starts_with("file://")

src/exec/qwen2_5vl.rs

@@ -19,7 +19,9 @@ impl ExecModel for Qwen2_5VLExec {
         } else {
             input_text.clone()
         };
-        let url = &input[1];
+        let url = input.get(1).ok_or_else(|| {
+            anyhow::anyhow!("qwen2.5vl requires a second input: image path or URL")
+        })?;
         let input_url = if url.starts_with("http://")
             || url.starts_with("https://")
             || url.starts_with("file://")

src/models/mod.rs

@@ -98,14 +98,16 @@ impl<'a> GenerateModel for ModelInstance<'a> {
             ModelInstance::Lfm2VL(model) => model.generate(mes),
             ModelInstance::Qwen2_5VL(model) => model.generate(mes),
             ModelInstance::Qwen3(model) => model.generate(mes),
-            ModelInstance::Qwen3_5(model) => model.generate(mes),
+
             ModelInstance::Qwen3Embedding(_) => {
                 Err(anyhow!("embedding model does not support chat completions"))
             }
-            ModelInstance::Qwen3ASR(model) => model.generate(mes),
             ModelInstance::Qwen3Reranker(_) => {
                 Err(anyhow!("reranker model does not support chat completions"))
             }
+            ModelInstance::Qwen3_5(model) => model.generate(mes),
+
+            ModelInstance::Qwen3ASR(model) => model.generate(mes),
             ModelInstance::Qwen3VL(model) => model.generate(mes),
             ModelInstance::DeepSeekOCR(model) => model.generate(mes),
             ModelInstance::HunyuanOCR(model) => model.generate(mes),
@@ -138,14 +140,15 @@ impl<'a> GenerateModel for ModelInstance<'a> {
             ModelInstance::Lfm2VL(model) => model.generate_stream(mes),
             ModelInstance::Qwen2_5VL(model) => model.generate_stream(mes),
             ModelInstance::Qwen3(model) => model.generate_stream(mes),
-            ModelInstance::Qwen3_5(model) => model.generate_stream(mes),
-            ModelInstance::Qwen3ASR(model) => model.generate_stream(mes),
+
             ModelInstance::Qwen3Embedding(_) => Err(anyhow!(
                 "embedding model does not support streaming chat completions"
             )),
-            ModelInstance::Qwen3Reranker(_) => {
-                Err(anyhow!("reranker model does not support chat completions"))
-            }
+            ModelInstance::Qwen3Reranker(_) => Err(anyhow!(
+                "reranker model does not support streaming chat completions"
+            )),
+            ModelInstance::Qwen3_5(model) => model.generate_stream(mes),
+            ModelInstance::Qwen3ASR(model) => model.generate_stream(mes),
             ModelInstance::Qwen3VL(model) => model.generate_stream(mes),
             ModelInstance::DeepSeekOCR(model) => model.generate_stream(mes),
             ModelInstance::HunyuanOCR(model) => model.generate_stream(mes),