refactor: Update file path handling in TopicSimilarityService for improved error reporting and ensure absolute paths for topic and embeddings cache files

config.py

@@ -12,7 +12,7 @@ load_dotenv()
 
 # ============ DIRECTORIES ============
 API_DIR = Path(__file__).parent
-PROJECT_ROOT = API_DIR.parent
+PROJECT_ROOT = API_DIR  # config.py is in the project root, so API_DIR is the project root
 
 # ============ HUGGING FACE MODELS ============
 HUGGINGFACE_API_KEY = os.getenv("HUGGINGFACE_API_KEY", "")

services/topic_similarity_service.py

@@ -62,15 +62,26 @@ class TopicSimilarityService:
     
     def _load_topics(self) -> List[str]:
         """Load topics from topics.json file"""
-        if not TOPICS_FILE.exists():
-            raise FileNotFoundError(f"Topics file not found: {TOPICS_FILE}")
+        # Ensure path is absolute
+        topics_file = Path(TOPICS_FILE).resolve()
+        
+        if not topics_file.exists():
+            raise FileNotFoundError(
+                f"Topics file not found: {topics_file}\n"
+                f"Current working directory: {Path.cwd()}\n"
+                f"PROJECT_ROOT: {PROJECT_ROOT}\n"
+                f"TOPICS_FILE path: {TOPICS_FILE}"
+            )
         
         try:
-            with open(TOPICS_FILE, 'r', encoding='utf-8') as f:
+            with open(topics_file, 'r', encoding='utf-8') as f:
                 data = json.load(f)
-            return data.get("topics", [])
+            topics = data.get("topics", [])
+            if not topics:
+                raise ValueError(f"No topics found in {topics_file}")
+            return topics
         except (json.JSONDecodeError, KeyError) as e:
-            raise ValueError(f"Error loading topics from {TOPICS_FILE}: {str(e)}")
+            raise ValueError(f"Error loading topics from {topics_file}: {str(e)}")
     
     def _get_topics_hash(self, topics: List[str]) -> str:
         """Generate a hash of the topics list to verify cache validity"""
@@ -79,11 +90,14 @@ class TopicSimilarityService:
     
     def _load_cached_embeddings(self) -> Optional[np.ndarray]:
         """Load cached topic embeddings if they exist and are valid"""
-        if not EMBEDDINGS_CACHE_FILE.exists():
+        # Ensure path is absolute
+        cache_file = Path(EMBEDDINGS_CACHE_FILE).resolve()
+        
+        if not cache_file.exists():
             return None
         
         try:
-            with open(EMBEDDINGS_CACHE_FILE, 'r', encoding='utf-8') as f:
+            with open(cache_file, 'r', encoding='utf-8') as f:
                 cache_data = json.load(f)
             
             # Verify cache is valid by checking topics hash
@@ -117,12 +131,13 @@ class TopicSimilarityService:
         }
         
         try:
-            # Ensure directory exists
-            EMBEDDINGS_CACHE_FILE.parent.mkdir(parents=True, exist_ok=True)
+            # Ensure path is absolute and directory exists
+            cache_file = Path(EMBEDDINGS_CACHE_FILE).resolve()
+            cache_file.parent.mkdir(parents=True, exist_ok=True)
             
-            with open(EMBEDDINGS_CACHE_FILE, 'w', encoding='utf-8') as f:
+            with open(cache_file, 'w', encoding='utf-8') as f:
                 json.dump(cache_data, f, indent=2)
-            logger.info(f"Cached {len(embeddings)} topic embeddings to {EMBEDDINGS_CACHE_FILE}")
+            logger.info(f"Cached {len(embeddings)} topic embeddings to {cache_file}")
         except Exception as e:
             logger.warning(f"Could not save cached embeddings: {e}")