Upload 2 files

agents.py

@@ -65,7 +65,7 @@ def extract_json_array(text: str):
     match = re.search(r"(\[\s*{.*?}\s*\])", text, re.DOTALL)
     if not match:
         # Try to find any JSON array in the text
-        match = re.search(r"\[.*?\]", text, re.DOTALL)
+        match = re.search(r"(\[.*?\])", text, re.DOTALL)
         if not match:
             return []
     
@@ -89,6 +89,11 @@ def extract_json_array(text: str):
                 json_str = re.sub(r',\s*]', ']', json_str)
                 # Fix unquoted keys
                 json_str = re.sub(r'(\w+):', r'"\1":', json_str)
+                # Fix extra quotes around objects
+                json_str = re.sub(r'"\s*{\s*"', '{"', json_str)
+                json_str = re.sub(r'"\s*}\s*"', '"}', json_str)
+                # Fix missing commas between objects
+                json_str = re.sub(r'"\s*}\s*{', '"},{', json_str)
                 return json.loads(json_str)
             except Exception as e3:
                 print("[extract_json_array] JSON fixing failed:", e3)
@@ -128,6 +133,11 @@ def safe_json_parse(content: str, fallback_value=None):
                 fixed_content = re.sub(r'(\w+):', r'"\1":', fixed_content)
                 # Fix single quotes to double quotes
                 fixed_content = fixed_content.replace("'", '"')
+                # Fix extra quotes around objects
+                fixed_content = re.sub(r'"\s*{\s*"', '{"', fixed_content)
+                fixed_content = re.sub(r'"\s*}\s*"', '"}', fixed_content)
+                # Fix missing commas between objects
+                fixed_content = re.sub(r'"\s*}\s*{', '"},{', fixed_content)
                 return json.loads(fixed_content)
             except Exception as e3:
                 print(f"[safe_json_parse] JSON fixing failed: {e3}")
@@ -203,6 +213,21 @@ async def extract_books_node(state):
             if manual_books:
                 books = manual_books
                 print("[extract_books_node] Manual extraction successful:", books)
+            else:
+                # Last resort: try to extract from the specific malformed pattern we saw
+                print("[extract_books_node] Attempting pattern-based extraction")
+                # Look for patterns like "title": "Book Name"
+                title_matches = re.findall(r'"title":\s*"([^"]+)"', content)
+                author_matches = re.findall(r'"author":\s*"([^"]+)"', content)
+                
+                if title_matches:
+                    for i, title in enumerate(title_matches):
+                        author = author_matches[i] if i < len(author_matches) else "Unknown"
+                        manual_books.append({"title": title, "author": author})
+                    
+                    if manual_books:
+                        books = manual_books
+                        print("[extract_books_node] Pattern-based extraction successful:", books)
         
         print("[extract_books_node] Parsed books:", books)