fix: formatting

Author: ryanolson

.gitignore

@@ -91,3 +91,10 @@ generated-values.yaml
 .build/
 **/.devcontainer/.env
 TensorRT-LLM
+
+
+# Ruler Generated Files
+/.cursor/instructions.md
+/.cursor/instructions.md.bak
+/CLAUDE.md
+/CLAUDE.md.bak

lib/llm/src/preprocessor.rs

@@ -454,7 +454,9 @@ impl OpenAIPreprocessor {
                     );
 
                     // Check if this response has a finish_reason
-                    let has_finish_reason = response.data.as_ref()
+                    let has_finish_reason = response
+                        .data
+                        .as_ref()
                         .map(|d| d.finish_reason.is_some())
                         .unwrap_or(false);
 
@@ -516,9 +518,10 @@ impl OpenAIPreprocessor {
                     Some((response, inner))
                 } else {
                     // Stream has ended - check if we need to send a usage chunk
-                    if inner.response_generator.is_usage_enabled() &&
-                       inner.finish_reason_sent &&
-                       !inner.usage_chunk_sent {
+                    if inner.response_generator.is_usage_enabled()
+                        && inner.finish_reason_sent
+                        && !inner.usage_chunk_sent
+                    {
                         inner.usage_chunk_sent = true;
 
                         // Create the final usage chunk

lib/llm/src/protocols/openai/chat_completions/delta.rs

@@ -20,7 +20,9 @@ impl NvCreateChatCompletionRequest {
     /// * [`DeltaGenerator`] configured with model name and response options.
     pub fn response_generator(&self, request_id: String) -> DeltaGenerator {
         let options = DeltaGeneratorOptions {
-            enable_usage: self.inner.stream_options
+            enable_usage: self
+                .inner
+                .stream_options
                 .as_ref()
                 .map(|opts| opts.include_usage)
                 .unwrap_or(false),
@@ -267,7 +269,7 @@ impl DeltaGenerator {
             model: self.model.clone(),
             system_fingerprint: self.system_fingerprint.clone(),
             choices,
-            usage: None,  // Always None for chunks with content/choices
+            usage: None, // Always None for chunks with content/choices
             service_tier: self.service_tier.clone(),
         }
     }
@@ -287,7 +289,7 @@ impl DeltaGenerator {
             created: self.created,
             model: self.model.clone(),
             system_fingerprint: self.system_fingerprint.clone(),
-            choices: vec![],  // Empty choices for usage-only chunk
+            choices: vec![], // Empty choices for usage-only chunk
             usage: Some(usage),
             service_tier: self.service_tier.clone(),
         }

lib/llm/src/protocols/openai/completions/delta.rs

@@ -9,7 +9,9 @@ impl NvCreateCompletionRequest {
     // inspect the request to extract options
     pub fn response_generator(&self, request_id: String) -> DeltaGenerator {
         let options = DeltaGeneratorOptions {
-            enable_usage: self.inner.stream_options
+            enable_usage: self
+                .inner
+                .stream_options
                 .as_ref()
                 .map(|opts| opts.include_usage)
                 .unwrap_or(false),
@@ -161,7 +163,7 @@ impl DeltaGenerator {
                 finish_reason,
                 logprobs,
             }],
-            usage: None,  // Always None for chunks with content/choices
+            usage: None, // Always None for chunks with content/choices
         };
 
         NvCreateCompletionResponse { inner }
@@ -182,7 +184,7 @@ impl DeltaGenerator {
             created: self.created,
             model: self.model.clone(),
             system_fingerprint: self.system_fingerprint.clone(),
-            choices: vec![],  // Empty choices for usage-only chunk
+            choices: vec![], // Empty choices for usage-only chunk
             usage: Some(usage),
         };
 

lib/llm/tests/test_streaming_usage.rs

@@ -1,6 +1,7 @@
 // SPDX-FileCopyrightText: Copyright (c) 2024-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 // SPDX-License-Identifier: Apache-2.0
 
+use async_trait::async_trait;
 use dynamo_async_openai::types::{
     ChatCompletionRequestMessage, ChatCompletionRequestUserMessage,
     ChatCompletionRequestUserMessageContent, ChatCompletionStreamOptions,
@@ -9,11 +10,10 @@ use dynamo_async_openai::types::{
 use dynamo_llm::preprocessor::OpenAIPreprocessor;
 use dynamo_llm::protocols::common::llm_backend::{BackendOutput, FinishReason};
 use dynamo_llm::protocols::openai::chat_completions::NvCreateChatCompletionRequest;
-use async_trait::async_trait;
 use dynamo_runtime::engine::{AsyncEngineContext, AsyncEngineStream};
 use dynamo_runtime::protocols::annotated::Annotated;
-use futures::stream;
 use futures::StreamExt;
+use futures::stream;
 use std::pin::Pin;
 use std::sync::Arc;
 use std::sync::atomic::{AtomicBool, Ordering};
@@ -115,7 +115,11 @@ fn create_mock_backend_stream(
         },
     ];
 
-    let stream = stream::iter(outputs.into_iter().map(|output| Annotated::from_data(output)));
+    let stream = stream::iter(
+        outputs
+            .into_iter()
+            .map(|output| Annotated::from_data(output)),
+    );
 
     use dynamo_runtime::engine::ResponseStream;
     ResponseStream::new(Box::pin(stream), ctx)
@@ -244,7 +248,10 @@ async fn test_streaming_with_usage_compliance() {
             usage.completion_tokens, 3,
             "Should have 3 completion tokens"
         );
-        assert_eq!(usage.prompt_tokens, 0, "Should have 0 prompt tokens (not set in test)");
+        assert_eq!(
+            usage.prompt_tokens, 0,
+            "Should have 0 prompt tokens (not set in test)"
+        );
         assert_eq!(
             usage.total_tokens, 3,
             "Total tokens should be prompt + completion"
@@ -289,4 +296,4 @@ async fn test_streaming_with_usage_false() {
             );
         }
     }
-}
+}