Update handler.py

handler.py

@@ -1,35 +1,50 @@
 from typing import Dict, List, Any
-import spacy
+import importlib.util
 
 class EndpointHandler:
     def __init__(self, path=""):
-        # This loads the model. 
-        # 'path' points to the repo storage, but since we will install 
-        # the model via requirements.txt, we load it by package name.
+        """
+        The __init__ method is called when starting the Endpoint.
+        We perform the imports and model loading here to match your logic.
+        """
+        # 1. Check if spaCy is installed (Your specific error handling)
+        if importlib.util.find_spec("spacy") is None:
+            raise RuntimeError(
+                "SpaCy is required but not installed. Install it with:\n"
+                ".\\.venv\\Scripts\\python -m pip install spacy\n"
+                "Then download the model:\n"
+                ".\\.venv\\Scripts\\python -m spacy download en_core_web_sm"
+            )
+
+        import spacy
+
+        # 2. Load the model (Your specific error handling)
         try:
+            # We load the model directly by name since it's installed via requirements.txt
             self.nlp = spacy.load("en_core_web_sm")
-        except OSError:
-            # Fallback: try loading from path if you uploaded model files directly
-            self.nlp = spacy.load(path)
+        except Exception as e:
+            raise RuntimeError(
+                "SpaCy model 'en_core_web_sm' is required but not available. "
+                "Install it with:\n"
+                ".\\.venv\\Scripts\\python -m spacy download en_core_web_sm"
+            ) from e
 
     def __call__(self, data: Dict[str, Any]) -> List[str]:
         """
-        Args:
-            data (:obj:): The payload of the request. 
-            Usually contains {"inputs": "Your text here"}
+        The __call__ method is called on every request.
         """
-        # 1. Get the raw text from the input JSON
+        # 1. Extract inputs
+        # The payload usually comes as {"inputs": "some text"}
         raw_text = data.pop("inputs", data)
-        
-        # Handle edge case if input is sent as a list
+
+        # Handle edge case where inputs might be a list
         if isinstance(raw_text, list):
             raw_text = raw_text[0]
 
-        # 2. Process with spaCy (The exact logic you provided)
+        # 2. Run your processing logic
         doc = self.nlp(raw_text)
         
-        # 3. Extract sentences (Your specific list comprehension)
+        # 3. Apply your specific list comprehension
         raw_sentences = [s.text.strip() for s in doc.sents if s.text.strip()]
-        
-        # 4. Return the list of strings
+
         return raw_sentences