ArturG9 commited on
Commit
5d24867
1 Parent(s): a049857

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +14 -9
app.py CHANGED
@@ -45,23 +45,28 @@ def create_retriever_from_chroma(vectorstore_path="docs/chroma/", search_type='m
45
 
46
 
47
  # Check if vectorstore exists
48
- #if os.path.exists(vectorstore_path) and os.listdir(vectorstore_path):
49
  # Load the existing vectorstore
50
- # vectorstore = Chroma(persist_directory=vectorstore_path,embedding_function=embeddings)
51
- #else:
 
 
52
  # Load documents from the specified data path
53
- loader = DirectoryLoader('./data/', glob="./*.txt", loader_cls=TextLoader)
54
- docs = loader.load()
55
- text_splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=chunk_overlap,separators=["\n \n \n", "\n \n", "\n1" , "(?<=\. )", " ", ""])
56
- split_docs = text_splitter.split_documents(docs)
 
 
 
57
 
58
 
59
 
60
  # Create the vectorstore
61
- vectorstore = Chroma.from_documents(
62
  documents=split_docs, embedding=embeddings, persist_directory=vectorstore_path
63
  )
64
-
65
 
66
  retriever=vectorstore.as_retriever(search_type = search_type, search_kwargs={"k": k})
67
 
 
45
 
46
 
47
  # Check if vectorstore exists
48
+ if os.path.exists(vectorstore_path) and os.listdir(vectorstore_path):
49
  # Load the existing vectorstore
50
+ st.write("Vector store exists and is loaded")
51
+ vectorstore = Chroma(persist_directory=vectorstore_path,embedding_function=embeddings)
52
+
53
+ else:
54
  # Load documents from the specified data path
55
+ st.write("Vector store doesnt exist and will be created now")
56
+ loader = DirectoryLoader('./data/', glob="./*.txt", loader_cls=TextLoader)
57
+ docs = loader.load()
58
+ st.write("Docs loaded")
59
+
60
+ text_splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=chunk_overlap,separators=["\n \n \n", "\n \n", "\n1" , "(?<=\. )", " ", ""])
61
+ split_docs = text_splitter.split_documents(docs)
62
 
63
 
64
 
65
  # Create the vectorstore
66
+ vectorstore = Chroma.from_documents(
67
  documents=split_docs, embedding=embeddings, persist_directory=vectorstore_path
68
  )
69
+ st.write("VectorStore is created")
70
 
71
  retriever=vectorstore.as_retriever(search_type = search_type, search_kwargs={"k": k})
72