update

tools/ai-review/clients/llm_client.py

@@ -346,7 +346,9 @@ class LLMClient:
             config: Provider-specific configuration.
         """
         if provider not in self.PROVIDERS:
-            raise ValueError(f"Unknown provider: {provider}. Available: {list(self.PROVIDERS.keys())}")
+            raise ValueError(
+                f"Unknown provider: {provider}. Available: {list(self.PROVIDERS.keys())}"
+            )
 
         self.provider_name = provider
         self.config = config or {}
@@ -405,6 +407,8 @@ class LLMClient:
 
         Handles markdown code blocks and preamble text.
         """
+        import re
+
         content = content.strip()
 
         # Attempt 1: direct parse
@@ -413,16 +417,24 @@ class LLMClient:
         except json.JSONDecodeError:
             pass
 
-        # Attempt 2: Extract from markdown code blocks
+        # Attempt 2: Extract from markdown code blocks (improved regex)
         if "```" in content:
-            # Find the JSON block
-            import re
-            match = re.search(r"```(?:json)?\s*([\s\S]*?)\s*```", content)
-            if match:
-                try:
-                    return json.loads(match.group(1))
-                except json.JSONDecodeError:
-                    pass
+            # Try multiple code block patterns
+            patterns = [
+                r"```json\s*\n([\s\S]*?)\n```",  # ```json with newlines
+                r"```json\s*([\s\S]*?)```",  # ```json without newlines
+                r"```\s*\n([\s\S]*?)\n```",  # ``` with newlines
+                r"```\s*([\s\S]*?)```",  # ``` without newlines
+            ]
+
+            for pattern in patterns:
+                match = re.search(pattern, content)
+                if match:
+                    try:
+                        json_str = match.group(1).strip()
+                        return json.loads(json_str)
+                    except json.JSONDecodeError:
+                        continue
 
         # Attempt 3: Find first { and last }
         try:
@@ -435,17 +447,22 @@ class LLMClient:
             pass
 
         # Attempt 4: Fix common JSON errors (comments, trailing commas)
-        # This is risky but helpful for LLM output
         try:
             # Remove comments
-            import re
             json_str = re.sub(r"//.*", "", content)
             json_str = re.sub(r"/\*[\s\S]*?\*/", "", json_str)
-            return json.loads(json_str)
-        except json.JSONDecodeError as e:
-            # If all attempts fail, raise an error with the content for debugging
-            snippet = content[:500] + "..." if len(content) > 500 else content
-            raise ValueError(f"Failed to parse JSON response: {e}. Raw content snippet: {snippet!r}")
+            # Try to extract JSON after cleaning
+            start = json_str.find("{")
+            end = json_str.rfind("}")
+            if start != -1 and end != -1:
+                json_str = json_str[start : end + 1]
+                return json.loads(json_str)
+        except json.JSONDecodeError:
+            pass
+
+        # If all attempts fail, raise an error with the content for debugging
+        snippet = content[:500] + "..." if len(content) > 500 else content
+        raise ValueError(f"Failed to parse JSON response: {snippet!r}")
 
     @classmethod
     def from_config(cls, config: dict) -> "LLMClient":
@@ -469,7 +486,9 @@ class LLMClient:
             }
         elif provider == "openrouter":
             provider_config = {
-                "model": config.get("model", {}).get("openrouter", "anthropic/claude-3.5-sonnet"),
+                "model": config.get("model", {}).get(
+                    "openrouter", "anthropic/claude-3.5-sonnet"
+                ),
                 "temperature": config.get("temperature", 0),
                 "max_tokens": config.get("max_tokens", 16000),
             }