The first step in our project is to load the PDF documents into the system and prepare them for further processing using the PyPDF library and Langchain's TextSplitter.
After loading and splitting the documents, the next step is to create a vector store using the Titan embedding model from AWS Bedrock for efficient similarity search.
This function reads all the PDF files from the data folder, splits them into chunks using the RecursiveCharacterTextSplitter, and returns the resulting documents.
In this function, we create an instance of a vector store which allows efficient retrieval of relevant information from the processed documents.
[
Collection
]
[
|
...
]