From aa20306eb919fad5601ecb2a9b1a0f430f0bffce Mon Sep 17 00:00:00 2001
From: samanthajmichael <samanthajmichael@gmail.com>
Date: Tue, 10 Dec 2024 22:34:19 -0500
Subject: [PATCH] reformatted

---
 frontend/core/__init__.py |  2 +-
 frontend/core/helpers.py  | 18 +++++++++++++++---
 frontend/core/rag.py      | 13 +++++++++----
 3 files changed, 25 insertions(+), 8 deletions(-)

diff --git a/frontend/core/__init__.py b/frontend/core/__init__.py
index d739db6..f7a5388 100644
--- a/frontend/core/__init__.py
+++ b/frontend/core/__init__.py
@@ -9,10 +9,10 @@
     process_uploaded_tc,
     read_file_content,
     read_pdf_to_string,
+    retrieve_all_metadata,
     retrieve_context_per_question,
     show_context,
     text_wrap,
-    retrieve_all_metadata
 )
 from .rag import (
     create_documents_with_metadata,
diff --git a/frontend/core/helpers.py b/frontend/core/helpers.py
index 0655d23..c68a9ba 100644
--- a/frontend/core/helpers.py
+++ b/frontend/core/helpers.py
@@ -258,7 +258,9 @@ def retrieve_all_metadata(vectorstore):
             titles = {doc.metadata.get("title", "Unknown") for doc in documents}
             return sorted(titles)
         else:
-            raise ValueError("Vectorstore does not have a valid 'docstore' or metadata.")
+            raise ValueError(
+                "Vectorstore does not have a valid 'docstore' or metadata."
+            )
     except Exception as e:
         raise ValueError(f"Metadata retrieval error: {e}")
 
@@ -274,7 +276,13 @@ def retrieve_context_per_question(question, retriever):
     Returns:
         list: A list of metadata titles if the query is about terms, or context otherwise.
     """
-    if any(keyword in question.lower() for keyword in ["what terms and conditions do you have access to", "what companies terms and conditions do you have access to"]):
+    if any(
+        keyword in question.lower()
+        for keyword in [
+            "what terms and conditions do you have access to",
+            "what companies terms and conditions do you have access to",
+        ]
+    ):
         try:
             return retrieve_all_metadata(retriever.vectorstore)
         except Exception as e:
@@ -283,7 +291,11 @@ def retrieve_context_per_question(question, retriever):
     # Retrieve relevant context for general questions
     try:
         results = retriever.get_relevant_documents(question)
-        return [doc.page_content for doc in results] if results else ["No relevant context found."]
+        return (
+            [doc.page_content for doc in results]
+            if results
+            else ["No relevant context found."]
+        )
     except Exception as e:
         raise ValueError(f"Error retrieving context: {e}")
 
diff --git a/frontend/core/rag.py b/frontend/core/rag.py
index d5bd4e6..2f91c73 100644
--- a/frontend/core/rag.py
+++ b/frontend/core/rag.py
@@ -33,7 +33,9 @@ def load_metadata(metadata_path):
         return json.load(f)
 
 
-def create_documents_with_metadata(metadata_list, data_folder, chunk_size=1000, chunk_overlap=200):
+def create_documents_with_metadata(
+    metadata_list, data_folder, chunk_size=1000, chunk_overlap=200
+):
     """
     Create documents with metadata and content chunks.
     """
@@ -58,9 +60,13 @@ def create_documents_with_metadata(metadata_list, data_folder, chunk_size=1000,
     return documents
 
 
-def initialize_vectorstore_with_metadata(metadata_file, data_folder, chunk_size=1000, chunk_overlap=200):
+def initialize_vectorstore_with_metadata(
+    metadata_file, data_folder, chunk_size=1000, chunk_overlap=200
+):
     metadata_list = load_metadata(metadata_file)
-    documents = create_documents_with_metadata(metadata_list, data_folder, chunk_size, chunk_overlap)
+    documents = create_documents_with_metadata(
+        metadata_list, data_folder, chunk_size, chunk_overlap
+    )
     embeddings = OpenAIEmbeddings()
 
     # Ensure metadata is being printed for debugging
@@ -99,7 +105,6 @@ def initialize_rag(metadata_file=metadata_path, data_folder=data_dir, k=2):
         raise ValueError(f"Failed to initialize RAG system: {e}")
 
 
-
 def encode_documents(path, chunk_size=1000, chunk_overlap=200):
     """
     Encodes all text files into a vector store using OpenAI embeddings and includes metadata.