Code review fixes

Author: jakozaur

src/generate_example.py

@@ -175,7 +175,7 @@ def create_reproduction_notebook() -> None:
                     "    print('='*50)\n",
                     "    \n",
                     "    # Get model response\n",
-                    "    response = client.generate(prompt, temperature=0.7)\n",
+                    "    response = client.generate(prompt)\n",
                     "    \n",
                     "    if response.error:\n",
                     '        print(f"❌ Error: {response.error}")\n',

src/utils/llm_backend.py

@@ -13,13 +13,14 @@ class LLMBackend(ABC):
     def __init__(self, config: dict[str, Any], seed: int | None = None) -> None:
         self.config = config
         self.seed = seed
+        # Set temperature based on seed - 0.0 for reproducibility, 0.7 otherwise
+        self.temperature = 0.0 if seed is not None else 0.7
 
     @abstractmethod
     def generate(
         self,
         prompt: str,
         system_prompt: str | None = None,
-        temperature: float = 0.7,
         max_tokens: int | None = None,
         stream: bool = False,
     ) -> ModelResponse:
@@ -30,7 +31,6 @@ def generate(
     def chat(
         self,
         messages: list[dict[str, str]],
-        temperature: float = 0.7,
         max_tokens: int | None = None,
     ) -> ModelResponse:
         """Multi-turn chat conversation."""
@@ -79,37 +79,27 @@ def generate(
         self,
         prompt: str,
         system_prompt: str | None = None,
-        temperature: float = 0.7,
         max_tokens: int | None = None,
         stream: bool = False,
     ) -> ModelResponse:
         """Generate response from Ollama model."""
-        # For reproducibility, use temperature=0 when seed is set
-        if self.seed is not None:
-            temperature = 0.0
-
         return self.client.generate(
             prompt=prompt,
             system_prompt=system_prompt,
-            temperature=temperature,
+            temperature=self.temperature,
             max_tokens=max_tokens,
             stream=stream,
         )
 
     def chat(
         self,
         messages: list[dict[str, str]],
-        temperature: float = 0.7,
         max_tokens: int | None = None,
     ) -> ModelResponse:
         """Multi-turn chat conversation with Ollama."""
-        # For reproducibility, use temperature=0 when seed is set
-        if self.seed is not None:
-            temperature = 0.0
-
         return self.client.chat(
             messages=messages,
-            temperature=temperature,
+            temperature=self.temperature,
             max_tokens=max_tokens,
         )
 
@@ -168,14 +158,10 @@ def generate(
         self,
         prompt: str,
         system_prompt: str | None = None,
-        temperature: float = 0.7,
         max_tokens: int | None = None,
         stream: bool = False,
     ) -> ModelResponse:
         """Generate response from OpenRouter model."""
-        # For reproducibility, use temperature=0 when seed is set
-        if self.seed is not None:
-            temperature = 0.0
 
         start_time = time.time()
 
@@ -189,7 +175,7 @@ def generate(
             request_params = {
                 "model": self.model,
                 "messages": messages,
-                "temperature": temperature,
+                "temperature": self.temperature,
                 "stream": stream,
                 "timeout": self.timeout,
                 "extra_headers": self._get_headers(),
@@ -238,13 +224,9 @@ def generate(
     def chat(
         self,
         messages: list[dict[str, str]],
-        temperature: float = 0.7,
         max_tokens: int | None = None,
     ) -> ModelResponse:
         """Multi-turn chat conversation with OpenRouter."""
-        # For reproducibility, use temperature=0 when seed is set
-        if self.seed is not None:
-            temperature = 0.0
 
         start_time = time.time()
 
@@ -253,7 +235,7 @@ def chat(
             request_params = {
                 "model": self.model,
                 "messages": messages,
-                "temperature": temperature,
+                "temperature": self.temperature,
                 "timeout": self.timeout,
                 "extra_headers": self._get_headers(),
             }

src/utils/model_client.py

@@ -38,6 +38,8 @@ def __init__(
         self.base_url = f"http://{host}:{port}"
         self.model = model
         self.seed = seed
+        # Set temperature based on seed - 0.0 for reproducibility, 0.7 otherwise
+        self.temperature = 0.0 if seed is not None else 0.7
 
     def _make_request(
         self, endpoint: str, data: dict[str, Any] | None = None, method: str = "POST"
@@ -205,14 +207,10 @@ def generate(
         self,
         prompt: str,
         system_prompt: str | None = None,
-        temperature: float = 0.7,
         max_tokens: int | None = None,
         stream: bool = False,
     ) -> ModelResponse:
         """Generate response from model"""
-        # For reproducibility, use temperature=0 when seed is set
-        if self.seed is not None:
-            temperature = 0.0
 
         start_time = time.time()
 
@@ -221,7 +219,7 @@ def generate(
             "prompt": prompt,
             "stream": stream,
             "options": {
-                "temperature": temperature,
+                "temperature": self.temperature,
             },
         }
 
@@ -262,13 +260,9 @@ def generate(
     def chat(
         self,
         messages: list[dict[str, str]],
-        temperature: float = 0.7,
         max_tokens: int | None = None,
     ) -> ModelResponse:
         """Multi-turn chat conversation"""
-        # For reproducibility, use temperature=0 when seed is set
-        if self.seed is not None:
-            temperature = 0.0
 
         start_time = time.time()
 
@@ -277,7 +271,7 @@ def chat(
             "messages": messages,
             "stream": False,
             "options": {
-                "temperature": temperature,
+                "temperature": self.temperature,
             },
         }