Create app.py

app.py

@@ -0,0 +1,44 @@
+from langchain.document_loaders import PyPDFDirectoryLoader
+
+loader = PyPDFDirectoryLoader("./data")
+docs = loader.load()
+
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+
+text_splitter = RecursiveCharacterTextSplitter(
+    chunk_size = 100, 
+    chunk_overlap  = 20, 
+    length_function = len, 
+)
+
+documents = text_splitter.split_documents(docs)
+
+import os
+os.environ["OPENAI_API_KEY"] = "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"
+
+from langchain.embeddings.openai import OpenAIEmbeddings
+embeddings = OpenAIEmbeddings() 
+
+from langchain.vectorstores import Chroma
+persist_directory = "vector_db"
+vectordb = Chroma.from_documents(docs, embedding = embeddings, persist_directory=persist_directory)
+
+vectordb.persist()
+vectordb = None
+
+# As you can see when you run the following cell - 
+# loaded the persisted vectore store is much quicker than reinstantiating it - 
+# and that is the benefit of persist_directory!
+vectordb = Chroma(persist_directory=persist_directory, embedding_function=embeddings)
+
+# Bring up ChatOpenAI
+from langchain.chat_models import ChatOpenAI
+llm = ChatOpenAI(model_name = "gpt-3.5-turbo")
+
+# Set up the document vector store as a Retriever tool
+doc_retriever = vectordb.as_retriever()
+
+# Now setup the RetrievalQA chain and leverage all the documents in the Vector DB
+from langchain.chains import RetrievalQA
+EN_17272_qa = RetrievalQA.from_chain_type(llm, chain_type="stuff", retriever=doc_retriever)
+