Minor changes to nomenclature

1 files changed, 8 insertions, 8 deletions

diff --git a/pdf_reader.py b/pdf_reader.py
index af7b3d1..e70a88b 100644
--- a/pdf_reader.py
+++ b/pdf_reader.py
@@ -92,21 +92,22 @@ def save_metadata(persist_directory, metadata):
         json.dump(metadata, f)
 
 
-def process_pdf_for_qa(pdf_path, embeddings, persist_directory):
-    """Prepare a PDF for question answering, using Chroma's persistence."""
+def process_pdf_for_qa(pdf_path, embeddings, base_persist_directory):
+    """Prepare a PDF for question answering, using a unique Chroma persistence directory for each PDF."""
 
     pdf_hash = compute_pdf_hash(pdf_path)
+    persist_directory = os.path.join(base_persist_directory, pdf_hash)  # Use hash to create a unique directory
 
     # Load or initialize metadata
-    metadata = load_metadata(persist_directory)
+    metadata = load_metadata(base_persist_directory)
 
     # Check if this PDF has already been processed
     if pdf_hash in metadata['processed_pdfs']:
-        print("This PDF has already been processed. Loading embeddings from the database.")
+        print(f"This PDF has already been processed. Loading embeddings from the database at {persist_directory}.")
         vector_store = Chroma(persist_directory=persist_directory, embedding_function=embeddings)
         return vector_store
 
-    # Initialize or load Chroma vector store
+    # Initialize or load Chroma vector store for this specific PDF
     vector_store = Chroma(persist_directory=persist_directory, embedding_function=embeddings)
 
     # Get text from PDF
@@ -130,7 +131,7 @@ def process_pdf_for_qa(pdf_path, embeddings, persist_directory):
 
     # Update the metadata to include this processed PDF hash
     metadata['processed_pdfs'].append(pdf_hash)
-    save_metadata(persist_directory, metadata)
+    save_metadata(base_persist_directory, metadata)
 
     return vector_store
 
@@ -165,7 +166,7 @@ def main():
     # Parse command line arguments
     parser = argparse.ArgumentParser(description='PDF Question Answering using LangChain and Ollama')
     parser.add_argument('pdf_path', type=str, help='Path to the PDF file')
-    parser.add_argument('--persist', type=str, default='db', help='Directory to save or load persisted vector store')
+    parser.add_argument('--persist', type=str, default='db', help='Base directory to save or load persisted vector stores')
     args = parser.parse_args()
 
     # Check if the PDF file exists
@@ -212,6 +213,5 @@ def main():
         else:
             print("No relevant information found for your question. Please try asking a different question.")
 
-
 if __name__ == "__main__":
     main()