Update src/ingestion/document_processor.py

src/ingestion/document_processor.py

@@ -8,19 +8,22 @@ class DocumentProcessor:
         self.embeddings = embeddings
 
     def process_pdfs(self, pdf_paths):
-        try:
-            documents = []
-            for path in pdf_paths:
-                loader = PyPDFLoader(path)
-                documents.extend(loader.load())
+        documents = []
 
-            splitter = RecursiveCharacterTextSplitter(
-                chunk_size=Config.CHUNK_SIZE,
-                chunk_overlap=Config.CHUNK_OVERLAP
-            )
-            splits = splitter.split_documents(documents)
+        for path in pdf_paths:
+            loader = PyPDFLoader(path)
+            docs = loader.load()
 
-            return FAISS.from_documents(splits, self.embeddings)
+            # Ensure source filename exists
+            for d in docs:
+                d.metadata["source"] = path.split("/")[-1]
 
-        except Exception as e:
-            raise RuntimeError(f"Document processing failed: {e}")
+            documents.extend(docs)
+
+        splitter = RecursiveCharacterTextSplitter(
+            chunk_size=Config.CHUNK_SIZE,
+            chunk_overlap=Config.CHUNK_OVERLAP
+        )
+
+        splits = splitter.split_documents(documents)
+        return FAISS.from_documents(splits, self.embeddings)